| [2010/02/08 00:01:29] @ Log started by gepetto | ||
| [2010/02/08 00:01:29] @ pinoyskullv2 is now known as pinoyskull | ||
| [2010/02/08 00:28:39] @ OpenMedia left channel #puppet () | ||
| [2010/02/08 00:39:02] @ Quit: themurph: Quit: themurph | ||
| [2010/02/08 00:40:08] @ themurph joined channel #puppet | ||
| [2010/02/08 00:53:19] <jamesturnbull> nevyn: that's never stopped any govt | ||
| [2010/02/08 00:53:37] <jamesturnbull> nevyn: give you $100 if DSD are compliant with the Telcoms Interception Act | ||
| [2010/02/08 01:11:29] @ LinuxCode joined channel #puppet | ||
| [2010/02/08 01:14:50] @ Quit: themurph: Quit: themurph | ||
| [2010/02/08 01:29:06] @ Quit: LinuxCode: Ping timeout: 252 seconds | ||
| [2010/02/08 01:42:11] @ LinuxCode joined channel #puppet | ||
| [2010/02/08 01:42:33] @ Quit: LinuxCode: Remote host closed the connection | ||
| [2010/02/08 01:47:21] @ Quit: pugnacity: Remote host closed the connection | ||
| [2010/02/08 01:53:22] @ pugnacity joined channel #puppet | ||
| [2010/02/08 02:00:44] @ Quit: jfreeman: Remote host closed the connection | ||
| [2010/02/08 02:03:11] @ Quit: vzctl_: Ping timeout: 258 seconds | ||
| [2010/02/08 02:03:30] @ vzctl_ joined channel #puppet | ||
| [2010/02/08 02:04:33] @ Quit: kolla: Remote host closed the connection | ||
| [2010/02/08 02:04:37] @ lak joined channel #puppet | ||
| [2010/02/08 02:14:31] @ skiold joined channel #puppet | ||
| [2010/02/08 02:27:23] @ Djelibeybi joined channel #puppet | ||
| [2010/02/08 02:36:45] @ giskard joined channel #puppet | ||
| [2010/02/08 02:41:18] @ PsychoSid joined channel #puppet | ||
| [2010/02/08 02:41:38] @ Quit: hamish: Remote host closed the connection | ||
| [2010/02/08 02:41:43] @ hamish joined channel #puppet | ||
| [2010/02/08 02:44:51] @ Quit: Djelibeybi: Quit: Leaving | ||
| [2010/02/08 02:46:29] @ Quit: vzctl_: Read error: Operation timed out | ||
| [2010/02/08 02:46:55] <gepetto> ::trac:: Recipes/cron edited @ http://reductivelabs.com/trac/puppet/wiki/Recipes/cron?version=21 | ||
| [2010/02/08 02:48:08] @ vzctl_ joined channel #puppet | ||
| [2010/02/08 02:49:35] @ nexx joined channel #puppet | ||
| [2010/02/08 02:51:50] @ pacalm joined channel #puppet | ||
| [2010/02/08 02:54:27] @ Quit: elementai: Quit: Lost terminal | ||
| [2010/02/08 02:54:32] @ Quit: giskard: Remote host closed the connection | ||
| [2010/02/08 02:54:53] @ giskard joined channel #puppet | ||
| [2010/02/08 02:55:37] @ kolla joined channel #puppet | ||
| [2010/02/08 02:56:17] <ericlee> Hi, I read some error messages from the /var/log/messages. | ||
| [2010/02/08 02:56:19] <ericlee> Feb 8 23:52:28 client1 puppetd[1312]: Reopening log files | ||
| [2010/02/08 02:56:19] <ericlee> Feb 8 23:52:28 client1 puppetd[1312]: Starting Puppet client version 0.24.8 | ||
| [2010/02/08 02:56:19] <ericlee> Feb 8 23:52:32 client1 puppetd[1312]: Could not call puppetmaster.getconfig: #<Errno::EHOSTUNREACH: No route to host - connect(2)> | ||
| [2010/02/08 02:56:20] <ericlee> Feb 8 23:52:32 client1 puppetd[1312]: Could not retrieve catalog: No route to host - connect(2) | ||
| [2010/02/08 02:56:24] <ericlee> What happened? | ||
| [2010/02/08 02:57:25] <gepetto> ::trac:: Recipes/cron edited @ http://reductivelabs.com/trac/puppet/wiki/Recipes/cron?version=22 | ||
| [2010/02/08 02:58:44] @ Quit: lak: Quit: lak | ||
| [2010/02/08 02:58:46] <ericlee> And when I run puppetd --test I encountered such error messages: err: Could not call puppetmaster.getconfig: #<Errno::EHOSTUNREACH: No route to host - connect(2)> | ||
| [2010/02/08 02:58:46] <ericlee> err: Could not retrieve catalog: No route to host - connect(2) | ||
| [2010/02/08 02:58:46] <ericlee> warning: Not using cache on failed catalog | ||
| [2010/02/08 02:58:55] <ericlee> What happened on earth? | ||
| [2010/02/08 02:59:30] @ Quit: giskard: Ping timeout: 248 seconds | ||
| [2010/02/08 02:59:59] <monachus> ericlee: that's a system message - is your puppetmaster reachable? is the name spelled correctly? does ping work? | ||
| [2010/02/08 03:00:35] <ericlee> I ping puppet (the puppetmaster) from client is okay. | ||
| [2010/02/08 03:01:06] @ Quit: Telmo: Ping timeout: 248 seconds | ||
| [2010/02/08 03:01:24] <ericlee> and puppetmaster is running/.\ | ||
| [2010/02/08 03:02:20] <monachus> well - it's not a puppet error message. it means that the client is unable to reach the master for some reason at the IP level - check your configs for typos or other errors. u can run with --debug and --verbose to see if it gives you more information. | ||
| [2010/02/08 03:02:41] <ericlee> It's not stable. | ||
| [2010/02/08 03:03:44] <ericlee> monachus, I set up master as puppet hostname. | ||
| [2010/02/08 03:03:53] <ericlee> No. | ||
| [2010/02/08 03:04:01] <ericlee> I mean puppet is the master hostname. | ||
| [2010/02/08 03:05:19] <ericlee> odd, ping is okay for resolve hostnames. | ||
| [2010/02/08 03:05:56] <monachus> i believe you - but the error message isn't lying. no route to host means that it is unable to reach whatever name/IP you've told it to. the most likely reason is a typo in a config - telling it to talk to pupppet or puppettt or something. if you can ping "puppet" from the CLI, look for something in your config, or in puppet.conf that is telling the puppet client to talk to something else. | ||
| [2010/02/08 03:06:36] @ elementai joined channel #puppet | ||
| [2010/02/08 03:07:37] <ericlee> I start puppet daemon at puppet (master) node, I find log said : | ||
| [2010/02/08 03:07:38] <ericlee> Feb 9 00:05:58 puppet puppetmasterd[2273]: Could not retrieve fact fqdn | ||
| [2010/02/08 03:07:38] <ericlee> Feb 9 00:05:58 puppet puppetmasterd[2273]: Host is missing hostname and/or domain: puppet | ||
| [2010/02/08 03:07:38] <ericlee> Feb 9 00:05:58 puppet puppetmasterd[2273]: Host is missing hostname and/or domain: puppet | ||
| [2010/02/08 03:07:38] <ericlee> Feb 9 00:05:58 puppet puppetmasterd[2273]: Could not find default node or by name with 'puppet' on node puppet | ||
| [2010/02/08 03:07:38] <ericlee> Feb 9 00:05:58 puppet puppetmasterd[2273]: Could not find default node or by name with 'puppet' on node puppet | ||
| [2010/02/08 03:07:41] <ericlee> Feb 9 00:05:58 puppet puppetd[2359]: Could not retrieve catalog: Could not find default node or by name with 'puppet' on node puppet | ||
| [2010/02/08 03:07:48] <ericlee> Why? | ||
| [2010/02/08 03:08:36] <monachus> do you have a node in your nodes.pp (or similar) file named "puppet" ? | ||
| [2010/02/08 03:09:48] @ poison joined channel #puppet | ||
| [2010/02/08 03:10:01] <gepetto> ::trac:: Release Notes edited @ http://reductivelabs.com/trac/puppet/wiki/ReleaseNotes?version=57 | ||
| [2010/02/08 03:11:36] <ericlee> |-- fileserver.conf | ||
| [2010/02/08 03:11:36] <ericlee> |-- manifests | ||
| [2010/02/08 03:11:36] <ericlee> | |-- classes | ||
| [2010/02/08 03:11:36] <ericlee> | | `-- hosts.pp | ||
| [2010/02/08 03:11:36] <ericlee> | |-- nodes | ||
| [2010/02/08 03:11:37] <ericlee> | | `-- client1.pp | ||
| [2010/02/08 03:11:39] <ericlee> | `-- site.pp | ||
| [2010/02/08 03:11:41] <ericlee> `-- puppet.conf | ||
| [2010/02/08 03:12:18] <monachus> nice graphic! | ||
| [2010/02/08 03:12:29] <nico> monachus: tree | ||
| [2010/02/08 03:12:32] <ericlee> yes | ||
| [2010/02/08 03:12:42] <monachus> yes - tree. i'm still waking up. | ||
| [2010/02/08 03:13:35] <ericlee> monachus, how do I know what client call the "puppet master" node? | ||
| [2010/02/08 03:14:21] <monachus> alright - this isn't a standard puppet tree - in order for this to work you'll have to have site.pp loading content from the other directories. have you read the design recommendations in wiki:PuppetBestPractice ? | ||
| [2010/02/08 03:14:22] <gepetto> monachus: wiki:PuppetBestPractice is http://reductivelabs.com/trac/puppet/wiki/PuppetBestPractice | ||
| [2010/02/08 03:14:40] <monachus> nodes connect to the puppetmaster and identify themselves with their fqdn | ||
| [2010/02/08 03:15:25] @ giskard joined channel #puppet | ||
| [2010/02/08 03:16:00] <monachus> it sounds like you have several things preventing your setup from working - client is unable to reach master (typo?), master is unable to identify client when it connects (tree?/site.pp?) and master is probably unable to identify anything in its config unless you have site.pp doing some stuff to load it. all of those items are covered in the documentation. | ||
| [2010/02/08 03:16:39] <ericlee> I feel puppet like ipchains. | ||
| [2010/02/08 03:16:58] <monachus> not at all. it's much friendlier than that. | ||
| [2010/02/08 03:18:18] <monachus> my recommendation would be to make a copy of your current config, go through the best practices guide to get a feel for what the puppetmaster wants to find, set up a basic first node, make sure that client can connect, and then use that as your foundation for building more recipes. | ||
| [2010/02/08 03:18:45] <monachus> after you understand how puppetmaster works, you can adapt the config to use your tree, if you want | ||
| [2010/02/08 03:19:06] <ericlee> but what's the default name of master? and how the client know the name??? | ||
| [2010/02/08 03:19:35] <monachus> it defaults to "puppet" | ||
| [2010/02/08 03:19:45] <ericlee> odd. | ||
| [2010/02/08 03:19:57] <monachus> unless you tell it to connect to something else with --server in puppet.conf or on the command line | ||
| [2010/02/08 03:20:11] <ericlee> bad conf rules. :( | ||
| [2010/02/08 03:21:09] @ ckauhaus joined channel #puppet | ||
| [2010/02/08 03:21:41] <ericlee> I just wanna put my resolv.conf to every machines, do I need all of them?> | ||
| [2010/02/08 03:22:01] <ericlee> And I think my folder structure is okay. | ||
| [2010/02/08 03:22:09] <ericlee> That makes me confused. | ||
| [2010/02/08 03:23:45] <ericlee> so do you have any idea? | ||
| [2010/02/08 03:34:39] <ericlee> Anyone here? | ||
| [2010/02/08 03:37:16] <nico> yes | ||
| [2010/02/08 03:37:35] <nico> you wanna "put your resolv.conf" ? what do you mean | ||
| [2010/02/08 03:39:30] <ericlee> Copy resolv.conf from /var/lib/puppet/files/etc/resolv.conf to other client machines. | ||
| [2010/02/08 03:40:27] <ericlee> I have restarted the puppet daemon in the client but nothing happened. | ||
| [2010/02/08 03:40:36] <ericlee> No error in logs. | ||
| [2010/02/08 03:40:40] <nico> stop here | ||
| [2010/02/08 03:40:43] @ Djelibeybi joined channel #puppet | ||
| [2010/02/08 03:40:47] <ericlee> wha? | ||
| [2010/02/08 03:40:58] <nico> first, create a manifest to distribute it | ||
| [2010/02/08 03:41:53] <ericlee> |-- fileserver.conf | ||
| [2010/02/08 03:41:53] <ericlee> |-- manifests | ||
| [2010/02/08 03:41:53] <ericlee> | |-- classes | ||
| [2010/02/08 03:41:53] <ericlee> | | `-- hosts.pp | ||
| [2010/02/08 03:41:53] <ericlee> | |-- nodes.pp | ||
| [2010/02/08 03:41:53] <ericlee> | `-- site.pp | ||
| [2010/02/08 03:41:55] <ericlee> `-- puppet.conf | ||
| [2010/02/08 03:41:57] <nico> pastie your fileserver.conf (of your master) to check params | ||
| [2010/02/08 03:42:13] <ericlee> [files] | ||
| [2010/02/08 03:42:14] <ericlee> path /var/lib/puppet/files | ||
| [2010/02/08 03:42:14] <ericlee> #deny *.evil.example.com | ||
| [2010/02/08 03:42:14] <ericlee> allow 192.168.1.0/24 | ||
| [2010/02/08 03:42:19] <nico> please, use pastie rather than brutal pastes in the channel | ||
| [2010/02/08 03:42:23] <nico> http://pastie.org/ | ||
| [2010/02/08 03:42:25] <ericlee> sorry | ||
| [2010/02/08 03:42:45] @ Quit: alban2: Ping timeout: 252 seconds | ||
| [2010/02/08 03:42:54] <nico> show your manifest that distributes resolv.cond | ||
| [2010/02/08 03:42:59] <nico> resolv.conf* | ||
| [2010/02/08 03:43:42] @ Quit: cmoates: Ping timeout: 272 seconds | ||
| [2010/02/08 03:43:44] <ericlee> http://pastebin.com/m7d387ecf | ||
| [2010/02/08 03:44:08] <nico> okay, looks fine | ||
| [2010/02/08 03:44:20] <nico> show your hosts definitions now | ||
| [2010/02/08 03:44:30] @ Quit: dyresen: Remote host closed the connection | ||
| [2010/02/08 03:47:24] <ericlee> http://pastebin.com/m2bda4bbd | ||
| [2010/02/08 03:47:35] @ dyresen joined channel #puppet | ||
| [2010/02/08 03:47:41] @ Quit: ppickfor: Ping timeout: 246 seconds | ||
| [2010/02/08 03:47:47] <ericlee> it's nodes.pp content. | ||
| [2010/02/08 03:48:24] <nico> ho I see | ||
| [2010/02/08 03:48:48] <nico> did you read the best practices ? your config tree is not right | ||
| [2010/02/08 03:49:23] <nico> you'd better wipe out what you've done and start from a white page | ||
| [2010/02/08 03:49:37] <nico> wiki:bestpractices | ||
| [2010/02/08 03:49:56] <ericlee> yes, I read that. | ||
| [2010/02/08 03:50:15] <ericlee> I just have one client now as a test. But I think nodes.pp I wrote is right, | ||
| [2010/02/08 03:50:26] @ gebi joined channel #puppet | ||
| [2010/02/08 03:50:40] <nico> is your client named client1 ? | ||
| [2010/02/08 03:50:44] <ericlee> yes | ||
| [2010/02/08 03:50:54] <nico> and ping client1 works ? | ||
| [2010/02/08 03:51:04] <ericlee> sure | ||
| [2010/02/08 03:51:26] <nico> and site.pp includes nodes.pp ? | ||
| [2010/02/08 03:51:26] <ericlee> and no error message in logs. | ||
| [2010/02/08 03:51:38] <ericlee> and site.pp includes nodes.pp ? What? | ||
| [2010/02/08 03:52:19] <ericlee> import "classes/*.pp" | ||
| [2010/02/08 03:52:32] <ericlee> just one line in site.pp | ||
| [2010/02/08 03:52:44] <nico> so nodes.pp is not imported | ||
| [2010/02/08 03:52:51] <nico> as it is not in classes | ||
| [2010/02/08 03:53:04] @ cmoates joined channel #puppet | ||
| [2010/02/08 03:53:07] <ericlee> oh damn. | ||
| [2010/02/08 03:53:10] <nico> there's no magic in puppet | ||
| [2010/02/08 03:53:19] <nico> only common sense | ||
| [2010/02/08 03:53:33] <ericlee> You know, I didn't find any good tutor in the puppet official site. | ||
| [2010/02/08 03:54:06] @ Quit: raphink: Remote host closed the connection | ||
| [2010/02/08 03:54:14] <Volcane> nico: ofcourse there's magic, its called modules and magically autoloads everything thats needed without the need for imports by giving you a convention for constructing and organizing related classes, templates and files | ||
| [2010/02/08 03:54:30] <ericlee> add import "nodes.pp" to site.pp? | ||
| [2010/02/08 03:55:00] <ericlee> official just says reference, but no tutor. junks. | ||
| [2010/02/08 03:57:39] @ raphink joined channel #puppet | ||
| [2010/02/08 03:58:20] <ericlee> Feb 9 00:57:42 puppet puppetmasterd[3355]: Could not parse for environment production: Syntax error at '{'; expected '}' at /etc/puppet/manifests/nodes.pp:2 | ||
| [2010/02/08 03:58:21] <ericlee> Feb 9 00:57:42 puppet puppetmasterd[3355]: Could not parse for environment production: Syntax error at '{'; expected '}' at /etc/puppet/manifests/nodes.pp:2 | ||
| [2010/02/08 03:58:28] <ericlee> I'm really dizzy. | ||
| [2010/02/08 03:59:14] @ danielbln joined channel #puppet | ||
| [2010/02/08 03:59:21] @ Quit: cmoates: Ping timeout: 245 seconds | ||
| [2010/02/08 03:59:33] <ericlee> http://pastebin.com/m2485fd1f <= nodes.pp | ||
| [2010/02/08 03:59:44] <ericlee> What's wrong with it? | ||
| [2010/02/08 04:00:08] <nico> Volcane: I don't call this magic, I call this reading the docs | ||
| [2010/02/08 04:00:17] @ alban2 joined channel #puppet | ||
| [2010/02/08 04:00:25] <monachus> ericlee: put { on same line as host "client1" | ||
| [2010/02/08 04:00:36] <ericlee> Why? Ruby rules? | ||
| [2010/02/08 04:00:42] <monachus> b/c that's the way it's done | ||
| [2010/02/08 04:01:14] <monachus> it has a parser which has reasonable expectations for what will be where. | ||
| [2010/02/08 04:01:56] <ericlee> oh | ||
| [2010/02/08 04:02:08] <ericlee> restrict format? | ||
| [2010/02/08 04:02:55] @ malraid joined channel #puppet | ||
| [2010/02/08 04:02:56] <ericlee> monachus, useless. | ||
| [2010/02/08 04:03:07] <monachus> ericlee: sorry i can't be of more help. | ||
| [2010/02/08 04:03:08] <ericlee> It shows Could not parse for environment production: Syntax error at '{'; expected '}' at /etc/puppet/manifests/nodes.pp:1 | ||
| [2010/02/08 04:03:15] @ ppickfor joined channel #puppet | ||
| [2010/02/08 04:03:18] <ericlee> the first line is wrong. | ||
| [2010/02/08 04:03:24] <monachus> ah - change "host" to "node" | ||
| [2010/02/08 04:03:32] <monachus> sorry - maybe it doesn't care where { is placed then. | ||
| [2010/02/08 04:03:43] <monachus> seriously, man - it's obvious that you haven't read the documentation | ||
| [2010/02/08 04:04:07] @ Quit: malraid: Client Quit | ||
| [2010/02/08 04:04:08] <monachus> we're here to help people who are stuck on something and have exhausted other means of assistance, not to provide free training to people who refuse to read the manual. | ||
| [2010/02/08 04:04:45] <ericlee> I read the manual. | ||
| [2010/02/08 04:04:56] <monachus> please read wiki:PuppetBestPractice and wiki:LanguageTutorial and set up an environment according to those documents. | ||
| [2010/02/08 04:04:57] <gepetto> monachus: wiki:PuppetBestPractice is http://reductivelabs.com/trac/puppet/wiki/PuppetBestPractice | ||
| [2010/02/08 04:05:00] <gepetto> monachus: wiki:LanguageTutorial is http://reductivelabs.com/trac/puppet/wiki/LanguageTutorial | ||
| [2010/02/08 04:05:25] <monachus> it's clear that you haven't, because you have host "client1" instead of node "client1" and you continue to insist that your configuration is correct when it won't parse, won't load, and won't run. | ||
| [2010/02/08 04:05:59] <ericlee> okay. | ||
| [2010/02/08 04:06:03] <ericlee> I understand. | ||
| [2010/02/08 04:06:07] <ericlee> My fault. | ||
| [2010/02/08 04:07:04] <monachus> puppet is awesome. it works. people use it to manage hundreds or thousands of systems. but it has a learning curve that is best processed through a slow start until you understand how it's dynamic nature can be tuned for your specific needs | ||
| [2010/02/08 04:07:33] <monachus> (damn - started a sentence w/ a conjunction and used it's instead of its. need breakfast.) | ||
| [2010/02/08 04:08:35] <ericlee> yep | ||
| [2010/02/08 04:11:04] @ cmoates joined channel #puppet | ||
| [2010/02/08 04:13:02] @ Quit: ppickfor: Ping timeout: 264 seconds | ||
| [2010/02/08 04:13:36] @ ppickfor joined channel #puppet | ||
| [2010/02/08 04:16:21] @ Robbie joined channel #puppet | ||
| [2010/02/08 04:16:49] @ Robbie is now known as Guest59333 | ||
| [2010/02/08 04:16:50] @ Quit: cmoates: Ping timeout: 248 seconds | ||
| [2010/02/08 04:19:23] @ fluxdude joined channel #puppet | ||
| [2010/02/08 04:20:40] @ mfladischer joined channel #puppet | ||
| [2010/02/08 04:23:56] @ TREllis joined channel #puppet | ||
| [2010/02/08 04:26:15] @ Mandus_ joined channel #puppet | ||
| [2010/02/08 04:27:37] @ verwilst joined channel #puppet | ||
| [2010/02/08 04:30:22] @ Quit: noyb: Quit: noyb | ||
| [2010/02/08 04:38:24] @ Quit: Djelibeybi: Quit: Leaving | ||
| [2010/02/08 04:41:49] @ Quit: ppickfor: Ping timeout: 264 seconds | ||
| [2010/02/08 04:42:48] @ ppickfor joined channel #puppet | ||
| [2010/02/08 04:44:19] @ cmoates joined channel #puppet | ||
| [2010/02/08 04:45:22] @ Welsh_Dwarf joined channel #puppet | ||
| [2010/02/08 04:46:08] <monachus> we've come across a problem recently that we're having difficulty tracking down | ||
| [2010/02/08 04:46:15] <monachus> all of our puppet clients end up with zombie cat processes: | ||
| [2010/02/08 04:46:16] <monachus> 27493 ? Z 0:00 [cat] <defunct> | ||
| [2010/02/08 04:46:25] <monachus> one per run, so after a few days, there are many | ||
| [2010/02/08 04:46:31] <monachus> all tied to the puppetd process | ||
| [2010/02/08 04:47:34] <monachus> we also just came across a machine with 131,000 files in /tmp with the output of "status" for some init script - thinking the two might be connected - does anyone know if puppet writes out "status" and then parses it with "cat" to determine process/service state? | ||
| [2010/02/08 04:48:18] <monachus> http://monachus.pastebin.com/d61c7d6f1 | ||
| [2010/02/08 04:49:22] @ Quit: cmoates: Ping timeout: 248 seconds | ||
| [2010/02/08 04:53:51] @ Ramonster joined channel #puppet | ||
| [2010/02/08 04:56:14] @ jab_doa joined channel #puppet | ||
| [2010/02/08 04:58:25] @ Quit: ppickfor: Ping timeout: 260 seconds | ||
| [2010/02/08 05:01:09] @ cmoates joined channel #puppet | ||
| [2010/02/08 05:06:26] @ Quit: cmoates: Ping timeout: 248 seconds | ||
| [2010/02/08 05:07:02] @ Quit: Mandus: Ping timeout: 256 seconds | ||
| [2010/02/08 05:13:53] @ henk joined channel #puppet | ||
| [2010/02/08 05:13:55] <henk> hi | ||
| [2010/02/08 05:14:13] <monachus> nm - those tmp files were from a busted heartbeat monitor | ||
| [2010/02/08 05:14:21] @ Quit: pinoyskull: Ping timeout: 245 seconds | ||
| [2010/02/08 05:14:21] @ ppickfor joined channel #puppet | ||
| [2010/02/08 05:14:42] <monachus> but the zombie cats are def from puppet | ||
| [2010/02/08 05:14:54] <monachus> henk: hi | ||
| [2010/02/08 05:16:53] @ ppickfor1 joined channel #puppet | ||
| [2010/02/08 05:16:54] @ Quit: ppickfor: Disconnected by services | ||
| [2010/02/08 05:16:54] @ maxagaz joined channel #puppet | ||
| [2010/02/08 05:16:55] <henk> i'm having the same problem i did have some time ago. i have several packages that require an mta. so i have this class mta: http://paste.debian.net/59102/ whenever a host uses postfix, i used to set '$mta = postfix'. now i changed some stuff and one of the postfix-nodes installed exim4... i created a node definition 'ourdefaults' including our standard classes and let every other node inherit from that. is that a bad idea? could that have broken the mta setup? | ||
| [2010/02/08 05:16:55] @ ppickfor1 is now known as ppickfor | ||
| [2010/02/08 05:17:49] @ cmoates joined channel #puppet | ||
| [2010/02/08 05:19:04] @ benlovell joined channel #puppet | ||
| [2010/02/08 05:23:30] @ Quit: cmoates: Ping timeout: 248 seconds | ||
| [2010/02/08 05:25:54] @ Quit: ppickfor: Ping timeout: 240 seconds | ||
| [2010/02/08 05:26:04] @ ppickfor joined channel #puppet | ||
| [2010/02/08 05:26:51] <monachus> henk: hmmm... | ||
| [2010/02/08 05:29:11] <monachus> tough to say without looking over everything. are you setting $mta = postfix in the node definition? | ||
| [2010/02/08 05:29:17] @ zazan joined channel #puppet | ||
| [2010/02/08 05:29:31] <henk> yes | ||
| [2010/02/08 05:29:35] <zazan> good morning | ||
| [2010/02/08 05:29:35] <monachus> but you include class mta in the ourdefaults node? | ||
| [2010/02/08 05:31:36] <henk> monachus: the "ourdefaults" node includes a "standardserver" class which includes other classes. so the class is included down to the "standardserver" class, which is included in the node "ourdefaults" which in turn is inherited by other nodes. | ||
| [2010/02/08 05:31:43] <zazan> guys, I'm using puppet 0.25.4 and I've been experiecing a bit of an issue. When not specifying any log destination puppet logs fine (/var/log/messages), on the other when specifying --logdest I can see in the newly created log file Puppet (notice): Reopening log files but it just stops there, if I want to see what actually puppet did afterwards I can only see it if I restart puppet | ||
| [2010/02/08 05:32:01] <monachus> henk: if you're acting on $mta in ourdefaults or anything included in ourdefaults, then that's the problem - puppet processes everything in the inherits before it processes the information in the node definition itself | ||
| [2010/02/08 05:32:37] <monachus> so if host1 includes ourdefaults, and if something included in ourdefaults looks at the value of $mta, $mta hasn't been set yet | ||
| [2010/02/08 05:32:48] <monachus> sorry - if host1 inherits ourdefaults | ||
| [2010/02/08 05:33:13] <henk> monachus: oh, that sounds likely... | ||
| [2010/02/08 05:33:23] <monachus> i ran into this trying to use tags | ||
| [2010/02/08 05:33:35] <monachus> and using a basenode with common classes included | ||
| [2010/02/08 05:33:51] <monachus> i ended up moving everything out of basenode and into base.pp which i include in every node definition | ||
| [2010/02/08 05:34:02] <monachus> so i can set tags and variables before loading any class that needs to act on them | ||
| [2010/02/08 05:36:34] <henk> monachus: ok, works perfectly without this extra inheritance layer. can you point me to some useful information regarding that base.pp? i never quite understood why there is site.pp, init.pp, base.pp, default.pp and whatever... | ||
| [2010/02/08 05:37:00] <monachus> henk: base.pp is just a class that i made up. let me put up a pastie for you | ||
| [2010/02/08 05:38:55] @ omry_|work joined channel #puppet | ||
| [2010/02/08 05:40:23] <ericlee> is there any cobbler and puppet combination docs? | ||
| [2010/02/08 05:41:21] @ Quit: ppickfor: Ping timeout: 260 seconds | ||
| [2010/02/08 05:42:21] @ toi joined channel #puppet | ||
| [2010/02/08 05:42:43] @ Quit: gebi: Read error: Operation timed out | ||
| [2010/02/08 05:42:52] <monachus> henk: http://monachus.pastebin.com/m237cbfac | ||
| [2010/02/08 05:43:08] <monachus> old and new ways - let me know if anything needs clarification | ||
| [2010/02/08 05:47:17] <henk> monachus: uhm, it really looks like just another class to me... | ||
| [2010/02/08 05:47:44] <monachus> yeah - that's all it is. but what matters is that it is included _after_ $mta is set, not inside of the basenode that node1 inherits from | ||
| [2010/02/08 05:47:58] <henk> monachus: but i guess i'm seeing to much in all those .pp files... | ||
| [2010/02/08 05:48:11] <henk> monachus: yeah, that's something worth noting it seems. | ||
| [2010/02/08 05:48:14] <monachus> i just named it base.pp - you can name it anything you want | ||
| [2010/02/08 05:48:24] <monachus> the inheritance isn't like class inheritance in the OO sense | ||
| [2010/02/08 05:48:36] <monachus> it's more like "go process everything up there and come back afterward" | ||
| [2010/02/08 05:48:42] <monachus> and puppet is all about what is processed when | ||
| [2010/02/08 05:51:04] @ cmoates joined channel #puppet | ||
| [2010/02/08 05:54:20] @ fruers joined channel #puppet | ||
| [2010/02/08 05:55:00] <fruers> Is it possible to "stack" variables? I want to do something like: | ||
| [2010/02/08 05:55:06] <fruers> $primary_nic="eth0" | ||
| [2010/02/08 05:55:21] <fruers> content => $network_${primary_nic} ? { | ||
| [2010/02/08 05:55:38] <fruers> Is it possible? | ||
| [2010/02/08 05:55:51] <zazan> :wq | ||
| [2010/02/08 05:55:54] <zazan> ops :) | ||
| [2010/02/08 05:56:34] @ Quit: cmoates: Ping timeout: 248 seconds | ||
| [2010/02/08 05:57:02] @ ppickfor joined channel #puppet | ||
| [2010/02/08 06:00:40] @ itguru joined channel #puppet | ||
| [2010/02/08 06:05:21] @ Quit: ppickfor: Disconnected by services | ||
| [2010/02/08 06:05:21] @ ppickfor1 joined channel #puppet | ||
| [2010/02/08 06:05:23] @ ppickfor1 is now known as ppickfor | ||
| [2010/02/08 06:05:31] <fruers> "${network_$nic}" give me ${network_wlan0} but it's not evaluating as a variable | ||
| [2010/02/08 06:05:38] @ hy is now known as _hy | ||
| [2010/02/08 06:06:59] @ _hy is now known as hy | ||
| [2010/02/08 06:07:49] @ cmoates joined channel #puppet | ||
| [2010/02/08 06:09:38] @ ffrank joined channel #puppet | ||
| [2010/02/08 06:10:30] <ffrank> hi. i want to set the group of a directory with recurse=true, but the mode for the directory only (non-recursively). will that work? | ||
| [2010/02/08 06:10:54] <ffrank> (or rather: how do i do this?) | ||
| [2010/02/08 06:10:54] @ Quit: toi: Read error: Connection reset by peer | ||
| [2010/02/08 06:11:32] @ toi joined channel #puppet | ||
| [2010/02/08 06:12:29] @ ahasenack joined channel #puppet | ||
| [2010/02/08 06:13:38] @ Quit: cmoates: Ping timeout: 248 seconds | ||
| [2010/02/08 06:14:51] <monachus> ffrank: mode is recursive as well, but puppet will add +x to a file mode for directory objects, so if you tell it mode => 0644 it will make files 0644 and directory 0755 | ||
| [2010/02/08 06:16:19] @ Quit: ppickfor: Ping timeout: 265 seconds | ||
| [2010/02/08 06:19:22] @ jcape joined channel #puppet | ||
| [2010/02/08 06:25:53] @ docelic__ joined channel #puppet | ||
| [2010/02/08 06:26:19] @ Quit: alexine_dsouza: Read error: Connection reset by peer | ||
| [2010/02/08 06:29:13] @ Quit: docelic_: Ping timeout: 264 seconds | ||
| [2010/02/08 06:29:50] <ffrank> monachus: ooh, that's good to know. thanks | ||
| [2010/02/08 06:29:51] @ Quit: omry|work: Quit: Leaving | ||
| [2010/02/08 06:30:07] <monachus> ffrank: np. | ||
| [2010/02/08 06:30:28] @ ppickfor joined channel #puppet | ||
| [2010/02/08 06:36:08] @ Quit: mfladischer: Ping timeout: 258 seconds | ||
| [2010/02/08 06:36:43] @ Quit: fruers: Quit: Leaving | ||
| [2010/02/08 06:36:51] @ mfladischer joined channel #puppet | ||
| [2010/02/08 06:40:35] @ Quit: raphink: Remote host closed the connection | ||
| [2010/02/08 06:40:54] @ cmoates joined channel #puppet | ||
| [2010/02/08 06:57:38] @ rmiller4pi8 joined channel #puppet | ||
| [2010/02/08 07:00:41] @ Quit: rmiller4pi81: Ping timeout: 260 seconds | ||
| [2010/02/08 07:09:33] @ mikepea joined channel #puppet | ||
| [2010/02/08 07:22:51] @ Pupeno joined channel #puppet | ||
| [2010/02/08 07:23:09] <Pupeno> keHow do you m one resource depend on another? | ||
| [2010/02/08 07:25:34] <henk> Pupeno: require => otherresource | ||
| [2010/02/08 07:25:57] <Pupeno> henk: thanks. | ||
| [2010/02/08 07:29:22] @ Quit: ppickfor: Ping timeout: 248 seconds | ||
| [2010/02/08 07:31:07] @ Quit: Mandus_: Quit: leaving | ||
| [2010/02/08 07:31:19] @ Mandus joined channel #puppet | ||
| [2010/02/08 07:39:53] @ ppickfor joined channel #puppet | ||
| [2010/02/08 07:41:41] <abien> is there a smart & easy way to make sure that from a bunch of classes, the stuff in a specific class (bootstrap) is allways ensured to be in place before anything else gets processed? | ||
| [2010/02/08 07:42:09] <abien> without having to require it all over the place in the other modules/classes. | ||
| [2010/02/08 07:45:16] <Pupeno> require => Package["ruby"] ? Is that the syntax? | ||
| [2010/02/08 07:45:36] <ptman> Pupeno, yes, it is | ||
| [2010/02/08 07:45:37] <henk> Pupeno: yes. | ||
| [2010/02/08 07:45:57] @ Quit: jcape: Ping timeout: 260 seconds | ||
| [2010/02/08 07:49:06] @ brothers joined channel #puppet | ||
| [2010/02/08 07:49:53] @ jcape joined channel #puppet | ||
| [2010/02/08 07:55:29] @ shug joined channel #puppet | ||
| [2010/02/08 07:56:35] @ suchu joined channel #puppet | ||
| [2010/02/08 07:56:48] @ Quit: ppickfor: Read error: Connection reset by peer | ||
| [2010/02/08 07:57:02] @ Quit: jcape: Ping timeout: 258 seconds | ||
| [2010/02/08 07:58:04] @ rmiller4pi81 joined channel #puppet | ||
| [2010/02/08 07:58:52] @ themurph joined channel #puppet | ||
| [2010/02/08 07:59:14] @ Quit: rmiller4pi8: Ping timeout: 248 seconds | ||
| [2010/02/08 08:00:31] @ Quit: themurph: Client Quit | ||
| [2010/02/08 08:02:32] @ tyll joined channel #puppet | ||
| [2010/02/08 08:04:43] @ themurph joined channel #puppet | ||
| [2010/02/08 08:05:29] @ illsci joined channel #puppet | ||
| [2010/02/08 08:06:35] @ Quit: benlovell: Quit: benlovell | ||
| [2010/02/08 08:10:15] @ suso joined channel #puppet | ||
| [2010/02/08 08:14:07] @ ppickfor joined channel #puppet | ||
| [2010/02/08 08:16:13] @ qwebirc99364 joined channel #puppet | ||
| [2010/02/08 08:16:48] <qwebirc99364> is it possible to have a deifinition inside a module's init.pp and import that definition inside site.pp ? | ||
| [2010/02/08 08:21:35] @ Quit: themurph: Quit: themurph | ||
| [2010/02/08 08:21:43] @ benlovell joined channel #puppet | ||
| [2010/02/08 08:23:10] @ jcape joined channel #puppet | ||
| [2010/02/08 08:27:41] @ Quit: jcape: Ping timeout: 245 seconds | ||
| [2010/02/08 08:29:43] @ Quit: toi: Ping timeout: 265 seconds | ||
| [2010/02/08 08:30:42] @ toi joined channel #puppet | ||
| [2010/02/08 08:30:49] @ Quit: gee: Quit: Lost terminal | ||
| [2010/02/08 08:33:00] <suso> qwebirc99364: I was having the same problem - it seems you need to write the definition with the full qualified name for it to autoload | ||
| [2010/02/08 08:33:21] <suso> define module::definition | ||
| [2010/02/08 08:33:56] <suso> then you use it with the full name | ||
| [2010/02/08 08:34:16] <gepetto> ::trac:: Managed By Puppet edited @ http://reductivelabs.com/trac/puppet/wiki/ManagedByPuppet?version=31 | ||
| [2010/02/08 08:34:16] <gepetto> ::trac:: Managed By Puppet edited @ http://reductivelabs.com/trac/puppet/wiki/ManagedByPuppet?version=32 | ||
| [2010/02/08 08:34:21] <suso> (If you do "define definition" then call module::definition it doesnt work) | ||
| [2010/02/08 08:38:33] @ themurph joined channel #puppet | ||
| [2010/02/08 08:39:47] @ bitfield joined channel #puppet | ||
| [2010/02/08 08:39:55] <bitfield> hey hey | ||
| [2010/02/08 08:39:56] @ Quit: rgsteele: Ping timeout: 246 seconds | ||
| [2010/02/08 08:40:58] <qwebirc99364> suso: so for example in yum module inside init.pp i have "define updatePackages ($packageList) { package { $packageList: ensure => latest } }" to call it inside site.pp I should call it using import yum::updatePackages {.....} ? | ||
| [2010/02/08 08:41:42] <bitfield> has anyone got wiki ops? I tried to add a new recipe page but it wouldn't let me :/ | ||
| [2010/02/08 08:41:44] <suso> you shouldnt have to call import, if it is in a module it will import itself when you say yum::updatePackages(....) | ||
| [2010/02/08 08:41:54] @ Quit: cliff-hm: Ping timeout: 248 seconds | ||
| [2010/02/08 08:41:58] @ mattwynne joined channel #puppet | ||
| [2010/02/08 08:42:11] <qwebirc99364> suso: aha. I see | ||
| [2010/02/08 08:43:51] <suso> what I'm not sure (I'm a newbie :-P) is about writing the define. I had to make it full qualified too on the modules file, but I have the defines on separate files, not on the init. | ||
| [2010/02/08 08:44:02] @ Quit: itguru: Ping timeout: 248 seconds | ||
| [2010/02/08 08:45:06] @ Quit: ckauhaus: Quit: Leaving. | ||
| [2010/02/08 08:45:49] <qwebirc99364> suso: well even on trying the full qualified name i get the following error "puppetmasterd[32146]: Could not find resource type yum::updatepackages" | ||
| [2010/02/08 08:49:58] @ fcoury joined channel #puppet | ||
| [2010/02/08 08:50:12] <suso> dunno. Case? (updatepackages vs updatePackages?) | ||
| [2010/02/08 08:51:04] @ ckauhaus joined channel #puppet | ||
| [2010/02/08 08:52:11] <fcoury> I have 15 servers that needs to scp a file into a central location. Can anyone show me a good reference for automating adding the ssh keys from each of the servers (generating them if needed) into this central servers? | ||
| [2010/02/08 08:55:14] <zirpu> fcoury: check the facts for the servers stored on the puppetmaster. | ||
| [2010/02/08 08:55:39] <fcoury> OK, I will do that! | ||
| [2010/02/08 08:55:51] <fcoury> thank you zirpu | ||
| [2010/02/08 08:56:05] <shine> someone using puppet dashboard ? :) | ||
| [2010/02/08 08:56:24] <monachus> @qwebirc99364 call it like yum::updatepackages { $fqdn: } | ||
| [2010/02/08 08:57:21] <monachus> provided it's called "updatepackages" and is in a class called "yum" | ||
| [2010/02/08 08:58:23] <fcoury> if factsync enabled by default? | ||
| [2010/02/08 08:58:59] @ sts joined channel #puppet | ||
| [2010/02/08 08:59:02] <sts> hello folks. | ||
| [2010/02/08 09:00:15] <sts> i'm having problems getting puppetclient to work with puppetmaster. i changed my ssl cert CN and now i get several of the following log entries when running puppetclient: err: //Node[puppet01.site.net]/puppetserver/global/vim/File[/usr/share/vim/vimfiles]: Failed to generate additional resources during transaction: Cannot access mount[vim] | ||
| [2010/02/08 09:02:32] @ itguru joined channel #puppet | ||
| [2010/02/08 09:03:09] @ malraid joined channel #puppet | ||
| [2010/02/08 09:03:15] <sts> how can i fixt that? | ||
| [2010/02/08 09:03:57] <zirpu> you just regenerated the puppetmaster ssl cert due to the CN change? | ||
| [2010/02/08 09:04:13] <sts> zirpu: yes. | ||
| [2010/02/08 09:04:20] @ Quit: ffrank: Quit: leaving | ||
| [2010/02/08 09:04:36] @ Quit: bug: Quit: bug | ||
| [2010/02/08 09:05:06] <zirpu> check that the clients are connecting to the new CN instead of the old one. | ||
| [2010/02/08 09:05:49] @ Robbie_ joined channel #puppet | ||
| [2010/02/08 09:06:30] <qwebirc99364> monachus: thanks. that worked | ||
| [2010/02/08 09:07:40] <sts> zirpu: it's the same host (puppetmaster/puppetclient) atm. and i'm trying to run it with: puppetd -t --debug --server=puppet.site.net | ||
| [2010/02/08 09:07:48] @ Quit: Guest59333: Ping timeout: 252 seconds | ||
| [2010/02/08 09:07:59] <qwebirc99364> has anyone tried to structure their modules as described here http://serialized.net/2009/07/puppet-module-patterns/ | ||
| [2010/02/08 09:09:30] <monachus> qwebirc99364: based solely upon the fact that the person who posted that blog entry can't figure out that their quoted content shouldn't show " and > i wouldn't trust them to know what they're talking about. | ||
| [2010/02/08 09:09:48] <zirpu> sts: i'd check the /var/lib/puppet/ssl dir and see if the client is still using the old CN ssl cert it generated. | ||
| [2010/02/08 09:10:09] @ Quit: Robbie_: Read error: No route to host | ||
| [2010/02/08 09:10:10] @ Robbie__ joined channel #puppet | ||
| [2010/02/08 09:10:48] @ Quit: rmiller4pi81: Quit: Leaving. | ||
| [2010/02/08 09:11:00] <qwebirc99364> its unfortunate that modules dont allow hierachy. it would make declaring things so much cleaner. i.e. modules/services could contain all service config e.g. modules/services/yum/ and modules/datacenters could contain data-center specific settings e.g. modules/datacenter/ny/network/resolv.conf | ||
| [2010/02/08 09:12:06] <qwebirc99364> and then inside nodes.pp we could easily configure the server by just importing required modules from services and datanceters | ||
| [2010/02/08 09:12:16] <monachus> modules allow hierarchy through "inherits" and through "include" | ||
| [2010/02/08 09:12:33] <sts> zirpu: where in ssl/ are the client side certs - since puppetmaster is running on the same host, there is more than the client certs in here. | ||
| [2010/02/08 09:12:38] <monachus> we have a "module" called global, beneath which i have classes for everything we manage - http, heartbeat, drbd, tftp, applications, etc | ||
| [2010/02/08 09:12:56] <monachus> i then have "client" modules that have the same manifests as in global | ||
| [2010/02/08 09:12:59] <qwebirc99364> monachus: but that was all declared inside a single init.pp file right ? | ||
| [2010/02/08 09:13:28] <monachus> arces::http inherits global::http - all my defines are set in global::http, and i set some client-specific variables and then call the defines from arces::http, etc | ||
| [2010/02/08 09:14:40] <monachus> qwebirc99364: under the "modules" directory i have "global," "arces," "monachus," and other directories for clients, and beneath each of these i have manifests/ files/ templates/ that contain content. everything in a client manifest inherits from the global manifest | ||
| [2010/02/08 09:14:44] <qwebirc99364> i am looking more for nested module hierarchies | ||
| [2010/02/08 09:14:54] <monachus> wiki:PuppetBestPractice | ||
| [2010/02/08 09:14:55] <gepetto> monachus: wiki: wiki:PuppetBestPractice is http://reductivelabs.com/trac/puppet/wiki/PuppetBestPractice | ||
| [2010/02/08 09:15:41] <monachus> you can nest all you want - let's say you have modules/services/yum/init.pp - that would be loaded as "services::yum" | ||
| [2010/02/08 09:15:47] <zirpu> sts: check /var/lib/puppet/ssl/{certs,private_keys,public_keys} | ||
| [2010/02/08 09:16:19] <monachus> you could also have modules/services/yum.pp which would still be "services::yum" but would then need a modules/services/init.pp that says import "*" | ||
| [2010/02/08 09:16:25] <qwebirc99364> http://laeg.reductivelabs.com/trac/puppet/wiki/ModuleOrganisation under Naming it states that "While it might be desirable to allow module hierarchies, for now modules cannot be nested." | ||
| [2010/02/08 09:17:49] <monachus> that's ambiguous. what does "nested" mean? perhaps that you can't have "modules/services/yum/some/other/module" - but you can certainly have "modules/services/yum.pp" or "modules/services/yum/init.pp" | ||
| [2010/02/08 09:18:07] <monachus> both of which are services::yum | ||
| [2010/02/08 09:18:11] <qwebirc99364> ok thanks for explaning... | ||
| [2010/02/08 09:19:01] <qwebirc99364> correct me if i am wrong, if i have a definition inside yum module inside nodes.pp i would call it by doing services::yum::updatePackage {....} | ||
| [2010/02/08 09:19:43] <monachus> correct - that would call the define updatePackage() inside of the class services::yum | ||
| [2010/02/08 09:19:52] <monachus> class != module | ||
| [2010/02/08 09:20:08] <monachus> often they are named the same, but you can have service/yum.pp with several classes inside of it | ||
| [2010/02/08 09:21:45] <qwebirc99364> ok, need to do some tests with what I have learnt now | ||
| [2010/02/08 09:21:51] <qwebirc99364> thanks again monachus | ||
| [2010/02/08 09:22:04] <monachus> np! ;) | ||
| [2010/02/08 09:23:30] @ unxfrek joined channel #puppet | ||
| [2010/02/08 09:23:58] @ nico is now known as nyco | ||
| [2010/02/08 09:24:05] @ Quit: brothers: Remote host closed the connection | ||
| [2010/02/08 09:24:09] @ brothers joined channel #puppet | ||
| [2010/02/08 09:24:12] @ Quit: brothers: Client Quit | ||
| [2010/02/08 09:27:39] @ rmiller4pi81 joined channel #puppet | ||
| [2010/02/08 09:29:24] @ jbeez_ is now known as jbeez | ||
| [2010/02/08 09:30:17] @ ninjazjb joined channel #puppet | ||
| [2010/02/08 09:30:37] @ eventi joined channel #puppet | ||
| [2010/02/08 09:31:01] @ Bass10 joined channel #puppet | ||
| [2010/02/08 09:31:06] <eventi> What's this I read about lak coming to NYC? | ||
| [2010/02/08 09:31:32] @ fdz joined channel #puppet | ||
| [2010/02/08 09:36:02] @ nyco is now known as nico | ||
| [2010/02/08 09:36:28] @ cliff-hm joined channel #puppet | ||
| [2010/02/08 09:42:04] @ Quit: bitfield: Quit: Leaving. | ||
| [2010/02/08 09:51:32] @ Quit: PsychoSid: Quit: Leaving. | ||
| [2010/02/08 09:51:39] @ Quit: rmiller4pi81: Ping timeout: 258 seconds | ||
| [2010/02/08 09:56:28] @ joe-mac joined channel #puppet | ||
| [2010/02/08 09:57:58] @ rmiller4pi8 joined channel #puppet | ||
| [2010/02/08 10:00:11] @ Quit: rmiller4pi8: Client Quit | ||
| [2010/02/08 10:01:17] @ Quit: ckauhaus: Quit: Leaving. | ||
| [2010/02/08 10:01:51] @ Quit: tecto: Quit: tecto | ||
| [2010/02/08 10:05:55] @ pheezy joined channel #puppet | ||
| [2010/02/08 10:09:41] @ rmiller4pi8 joined channel #puppet | ||
| [2010/02/08 10:11:46] @ Quit: Robbie__: Read error: Connection reset by peer | ||
| [2010/02/08 10:12:11] @ Robbie_ joined channel #puppet | ||
| [2010/02/08 10:12:37] @ Quit: docelic__: Quit: Leaving | ||
| [2010/02/08 10:15:57] @ Quit: m1nish: Quit: Leaving | ||
| [2010/02/08 10:17:48] @ LukeL_ joined channel #puppet | ||
| [2010/02/08 10:18:35] @ brothers joined channel #puppet | ||
| [2010/02/08 10:20:04] @ ckauhaus joined channel #puppet | ||
| [2010/02/08 10:22:02] <ohadlevy> joe-mac: I'm guessing that you found out if puppet certs has a password? :) | ||
| [2010/02/08 10:22:21] @ rgsteele joined channel #puppet | ||
| [2010/02/08 10:22:41] @ labrown joined channel #puppet | ||
| [2010/02/08 10:24:30] <elasticdog> is there anything special you have to do to get the puppet master server to be a client to itself? | ||
| [2010/02/08 10:24:49] <joe-mac> ohadlevy: they don't but splunk still refuses to use them | ||
| [2010/02/08 10:24:57] <joe-mac> with a really helpful error | ||
| [2010/02/08 10:24:58] <henk> elasticdog: i found that calling your master server NOT puppetmaster is one of the things one should definitely do... | ||
| [2010/02/08 10:25:01] <joe-mac> "can't read cert" | ||
| [2010/02/08 10:25:10] <elasticdog> I keep getting "warning: Not using cache on failed catalog" "err: Could not retrieve catalog; skipping run" when doing puppetd --test | ||
| [2010/02/08 10:25:13] <joe-mac> yet openssl x509 reads tehm fine from both the root user and splunk user | ||
| [2010/02/08 10:25:17] <joe-mac> so annoying ohadlevy | ||
| [2010/02/08 10:25:30] <joe-mac> i want pki to be one single cert i don't want to have a bunch of certs per server, it's stupid | ||
| [2010/02/08 10:25:32] <ohadlevy> joe-mac: the only reason I referred you to func was because of a statement in their wiki | ||
| [2010/02/08 10:25:44] <joe-mac> what statement was that ohadlevy? i didn't catch it | ||
| [2010/02/08 10:25:59] <joe-mac> i finally just put splunk to use its own certs and it worked fine | ||
| [2010/02/08 10:26:04] <elasticdog> henk: :-), that's one of the main points drawing me in to using puppet | ||
| [2010/02/08 10:26:19] <ohadlevy> joe-mac: https://fedorahosted.org/func/wiki/FuncWithPuppet - search for passphrase | ||
| [2010/02/08 10:26:31] <ohadlevy> passphase | ||
| [2010/02/08 10:26:58] @ Quit: mfladischer: Quit: Ex-Chat | ||
| [2010/02/08 10:27:00] <ohadlevy> elasticdog: check your puppetmaster logs | ||
| [2010/02/08 10:27:51] <joe-mac> o holy shit | ||
| [2010/02/08 10:28:01] <joe-mac> ohadlevy: now my question is WTF is a passphase | ||
| [2010/02/08 10:28:06] <joe-mac> or is that just a typo | ||
| [2010/02/08 10:28:17] <ohadlevy> sounds like a typo | ||
| [2010/02/08 10:28:36] <ohadlevy> more than that, I'm using the certificates with apache, so I dont think apache knows that "passphase" :) | ||
| [2010/02/08 10:28:45] <joe-mac> siunce it's working now, i am not going to touch it, but dude, thanks for the info | ||
| [2010/02/08 10:28:49] <joe-mac> i was literally losing it | ||
| [2010/02/08 10:28:54] <elasticdog> ohadlevy: /var/log/httpd/access_log shows a GET from 127.0.0.1 with a big long cert looking string | ||
| [2010/02/08 10:29:15] <elasticdog> ohadlevy: (I'm using a passenger setup that works fine with other clients) | ||
| [2010/02/08 10:29:18] <ohadlevy> elasticdog: and the puppetmaster syslog? | ||
| [2010/02/08 10:29:27] <ohadlevy> joe-mac: no worries | ||
| [2010/02/08 10:29:46] <joe-mac> might have to move from cap to func for administration activities thoguh | ||
| [2010/02/08 10:29:55] <joe-mac> now that func is more mature especially | ||
| [2010/02/08 10:30:00] <ohadlevy> why? | ||
| [2010/02/08 10:30:02] <joe-mac> capistrano doesn't scale for shit | ||
| [2010/02/08 10:30:13] <joe-mac> well, net:ssh doesn't | ||
| [2010/02/08 10:30:17] <joe-mac> therefore, cap doesn't | ||
| [2010/02/08 10:30:42] <ohadlevy> joe-mac: yeah, I really tried using net:ssh | ||
| [2010/02/08 10:30:47] <ohadlevy> net:ssh:multi that is | ||
| [2010/02/08 10:31:01] <joe-mac> yea it's awful | ||
| [2010/02/08 10:31:02] <ohadlevy> but it hangs on 20+ servers | ||
| [2010/02/08 10:31:06] <joe-mac> scalability is ridiculouis | ||
| [2010/02/08 10:31:07] <joe-mac> yep | ||
| [2010/02/08 10:31:17] <joe-mac> likle i have 20 servers in my office, let alone the DC | ||
| [2010/02/08 10:31:25] <elasticdog> ohadlevy: don't have rsyslog installed, so no /var/log/syslog, just the apache stuff | ||
| [2010/02/08 10:31:46] <ohadlevy> I'm currently using some c based sw to fork ssh | ||
| [2010/02/08 10:31:57] <ohadlevy> but on the long term, mcollective like is much better | ||
| [2010/02/08 10:32:10] <SyTonnerre> My coworker has ~20 servers under his desk | ||
| [2010/02/08 10:32:24] <ohadlevy> elasticdog: so use puppetmaster in debug mode | ||
| [2010/02/08 10:32:38] @ bug joined channel #puppet | ||
| [2010/02/08 10:32:45] <ohadlevy> joe-mac: you can easily fork ssh yourself | ||
| [2010/02/08 10:33:10] <joe-mac> yea i don't have the programming chops to fix synchronization issues | ||
| [2010/02/08 10:33:34] <joe-mac> and yes, mcollective is the future of mass administration | ||
| [2010/02/08 10:33:35] <joe-mac> however | ||
| [2010/02/08 10:33:40] <joe-mac> it's overkill for what i have right now | ||
| [2010/02/08 10:34:17] <ohadlevy> joe-mac: what do you mean for sync issues? | ||
| [2010/02/08 10:34:18] @ tecto joined channel #puppet | ||
| [2010/02/08 10:34:19] @ Quit: tecto: Changing host | ||
| [2010/02/08 10:34:19] @ tecto joined channel #puppet | ||
| [2010/02/08 10:34:31] <nico> wiki:BestPractices | ||
| [2010/02/08 10:34:40] <elasticdog> ohadlevy: messages.log shows the same error messages "could not retrieve catalog, skipping run" | ||
| [2010/02/08 10:34:55] <joe-mac> i mean it's obviously a thread scaling issues | ||
| [2010/02/08 10:34:58] <joe-mac> concurrency i mean | ||
| [2010/02/08 10:35:02] @ allsystemsarego joined channel #puppet | ||
| [2010/02/08 10:35:04] @ jcape joined channel #puppet | ||
| [2010/02/08 10:35:04] <elasticdog> I'll try debug mode... | ||
| [2010/02/08 10:35:11] <joe-mac> past 20-30 boxes it starts failing in retarded ways | ||
| [2010/02/08 10:35:28] <joe-mac> sometimes | ||
| [2010/02/08 10:35:30] <joe-mac> this one is fun | ||
| [2010/02/08 10:35:36] <ohadlevy> joe-mac: just dont use ruby net:ssh :) | ||
| [2010/02/08 10:35:36] <joe-mac> sometimes, it wil duie, then my ssh agent will die with it | ||
| [2010/02/08 10:35:47] @ danbeck joined channel #puppet | ||
| [2010/02/08 10:35:48] <joe-mac> how it takes down my ssh agent itself is a mystery to me | ||
| [2010/02/08 10:35:49] <ohadlevy> joe-mac: yeah, thats another story | ||
| [2010/02/08 10:35:56] <joe-mac> it's infuriating | ||
| [2010/02/08 10:36:09] <joe-mac> i've had to actually like lock my box and walk away and go for a walk when that's happened to me | ||
| [2010/02/08 10:36:31] @ Quit: Robbie_: Remote host closed the connection | ||
| [2010/02/08 10:36:39] <rmiller4pi8> i know you guys aren't huge func fans, but I can't help but thing it would be less painful than a lot of these hacks | ||
| [2010/02/08 10:36:53] <joe-mac> what hacks? | ||
| [2010/02/08 10:36:54] @ Robbie_ joined channel #puppet | ||
| [2010/02/08 10:37:06] <joe-mac> fcking cofree on my other keyuboard | ||
| [2010/02/08 10:37:07] <joe-mac> nice | ||
| [2010/02/08 10:37:15] <joe-mac> AND on my passport app | ||
| [2010/02/08 10:37:17] <joe-mac> grrr | ||
| [2010/02/08 10:37:28] <rmiller4pi8> oopa | ||
| [2010/02/08 10:37:54] <joe-mac> luckily my linux trade show dolls are safe | ||
| [2010/02/08 10:38:21] <ohadlevy> rmiller4pI8: I guess that 80% of the problem is to get the right host, the rest is not suppose to be such a big deal | ||
| [2010/02/08 10:38:38] <ohadlevy> by the right host i mean the hosts to connect to | ||
| [2010/02/08 10:39:18] <rmiller4pi8> yeah, what i ended up doing is generating the func group configuration as a .erb template which uses activerecord to connect to the puppet database, and just enumerates the hosts that are in each puppet class | ||
| [2010/02/08 10:39:23] <joe-mac> o nice my w-2's got owned in the process :facepalm: mondays | ||
| [2010/02/08 10:39:32] <elasticdog> ohadlevy: stopping httpd and then using puppetmasterd --debug seemed to work...odd that the passenger stuff would make any difference | ||
| [2010/02/08 10:39:35] <joe-mac> rmiller4pi8: i do something similar with capistrano actually | ||
| [2010/02/08 10:39:48] <rmiller4pi8> yeah, i do that for func, cap, and nagios | ||
| [2010/02/08 10:39:49] <joe-mac> i conncet to the puppet store config db's and builds capistrano roles | ||
| [2010/02/08 10:40:14] <ohadlevy> rmiller4pi8: I also did something similar, you can chose hosts base on classes, fact, environment or state (e.g. puppet state - good, bad, errors, not running puppet etc) | ||
| [2010/02/08 10:40:22] <rmiller4pi8> i found that left me with more human-readable (hence debuggable) configs in those applications than exported resources | ||
| [2010/02/08 10:40:24] @ shenson joined channel #puppet | ||
| [2010/02/08 10:40:41] <rmiller4pi8> so then why all this ssh-wrapping stuff? | ||
| [2010/02/08 10:41:13] <joe-mac> what do you mean? | ||
| [2010/02/08 10:41:23] <joe-mac> capistrano uses ssh because it's a standard... | ||
| [2010/02/08 10:41:32] <ohadlevy> rmiller4pi8: the problem joe-mac is referring to is that most ssh cluster like sw, hangs when you have a lot of parallel sshs going on | ||
| [2010/02/08 10:41:41] @ kaptk2 joined channel #puppet | ||
| [2010/02/08 10:41:51] <rmiller4pi8> oh, so this problem is particular to using capistrano? | ||
| [2010/02/08 10:41:57] <rmiller4pi8> b/c it's already built around ssh? | ||
| [2010/02/08 10:41:59] <joe-mac> particular to ruby ssh | ||
| [2010/02/08 10:42:10] <joe-mac> yes, byut it being built around ssh is also its strength | ||
| [2010/02/08 10:42:20] <rmiller4pi8> sorry, i didn't realize you were describing a problem you were having specific to capistrano | ||
| [2010/02/08 10:42:31] <rmiller4pi8> that's what i get for walking in in the middle | ||
| [2010/02/08 10:42:47] <ohadlevy> its not specific to cap, | ||
| [2010/02/08 10:43:32] <Volcane> cap uses ruby threads to call its native ssh library - it doesnt spawn ssh | ||
| [2010/02/08 10:43:41] <Volcane> the problem is a i/o wait in one thread in ruby locks all other threads | ||
| [2010/02/08 10:43:43] <ohadlevy> usually you have a limit of how many parralel ssh connections you can do, ruby implement ion is probably one of the lowest... but usually you cant do more than 50 connections at the same time | ||
| [2010/02/08 10:43:47] <Volcane> and ruby threads arent native they're green threads | ||
| [2010/02/08 10:43:50] <Volcane> so it all just suck | ||
| [2010/02/08 10:45:06] <ohadlevy> volcane: yep | ||
| [2010/02/08 10:47:04] <joe-mac> yea there are lots of actual contributing factors to this one problem | ||
| [2010/02/08 10:47:38] * Volcane 's done 200 hosts concurrently without any even remote indication of anything being a problem with mcollective | ||
| [2010/02/08 10:47:54] @ Quit: tecto: Ping timeout: 252 seconds | ||
| [2010/02/08 10:48:41] <joe-mac> mcollective is overkill unless you reach a couple hundred hosts thoguh | ||
| [2010/02/08 10:48:46] <Pupeno> How do I run commands, such as passenger-install-apache2-module, which should be run after installing passenger and/or installing apache, or ufw commands to set up a firewall which should be run only once? | ||
| [2010/02/08 10:48:49] <Volcane> joe-mac: why? | ||
| [2010/02/08 10:48:52] <joe-mac> for instance you get about a 5 seconid wait generallyt right | ||
| [2010/02/08 10:49:00] <joe-mac> with cap i get a millisecond wait when i run commands | ||
| [2010/02/08 10:49:16] <joe-mac> because it's just ssh tunnels already established after tghe first connection | ||
| [2010/02/08 10:49:19] <Volcane> joe-mac: not really the initial wait is to discover your nodes, u can drop that way down on a fast network | ||
| [2010/02/08 10:49:33] <ohadlevy> i wonder if jruby have better threading support | ||
| [2010/02/08 10:49:41] <rmiller4pi8> ohadlevy: it does | ||
| [2010/02/08 10:49:42] <Volcane> ohadlevy: yeah those are native threads | ||
| [2010/02/08 10:50:20] <Volcane> joe-mac: the stuff like ping and fact reporting i dont bother discovering first so i just wait 5 seconds, but the timeouts are configurable | ||
| [2010/02/08 10:50:44] <joe-mac> Pupeno: you want to set them to refreshonly => "true" and only notify them when you need to | ||
| [2010/02/08 10:51:07] @ Quit: majeru: Read error: Operation timed out | ||
| [2010/02/08 10:51:16] <nasrat> or use creates with exec commands | ||
| [2010/02/08 10:51:25] <joe-mac> yea creates is a good one | ||
| [2010/02/08 10:51:54] <joe-mac> proabbly better in this scenario since it enabled some module which is a file i think | ||
| [2010/02/08 10:52:06] @ kami joined channel #puppet | ||
| [2010/02/08 10:52:55] <kami> hello | ||
| [2010/02/08 10:53:53] <Pupeno> joe-mac: I'm sorry, I'm really new to puppet and I lack context to what you say so I can't understand it. Can you point me to the documentation I should read? What type would the command be? | ||
| [2010/02/08 10:54:13] <ohadlevy> joe-mac: I dont think that func will be any faster then just forking ssh directly | ||
| [2010/02/08 10:54:13] <kami> I have an error saying "Duplicate definition: File[/opt/oracle/admin/] is already defined ..." | ||
| [2010/02/08 10:55:23] <kami> I've stared at the code for hours now and don't see any duplication. Is there a good strategy for such a case? | ||
| [2010/02/08 10:55:28] @ qwebirc24977 joined channel #puppet | ||
| [2010/02/08 10:55:58] @ Quit: qwebirc24977: Client Quit | ||
| [2010/02/08 10:56:47] @ Quit: suchu: Quit: ChatZilla 0.9.86 [Firefox 3.5.7/20091221164558] | ||
| [2010/02/08 10:56:50] <ohadlevy> kami: look closer :) it usually give you a line number | ||
| [2010/02/08 10:57:02] * Volcane guess its a define. | ||
| [2010/02/08 10:57:21] @ lilmatt joined channel #puppet | ||
| [2010/02/08 10:57:22] <rmiller4pi8> ohadlevy: no, it's not especially faster, it's just a nice interface that lets you define the level of parallelism you want, and you can pass objects and exceptions across the wire. obviously if you only have a few boxes it's not worth it--just use ssh. but if there's already pain around ssh and you're using some C program to fork it for you, i figure you've reached the point at which func is probably less hassle on net | ||
| [2010/02/08 10:58:11] <ohadlevy> rmiller4pi8: the main issue with func is that it works well only with RH/Fedora | ||
| [2010/02/08 10:58:25] <rmiller4pi8> ah, yes. that is very true. | ||
| [2010/02/08 10:58:27] <kami> ohadlevy: it gives me the same line number twice. And it happens only when I use one definition twice. So, the problem is definitely sitting in front of the computer. May I paste it somewhere? | ||
| [2010/02/08 10:58:29] <ohadlevy> and you can use bash fork too | ||
| [2010/02/08 10:58:41] <ohadlevy> kami: use pastie.org | ||
| [2010/02/08 10:58:49] <kami> ohadlevy: will do. | ||
| [2010/02/08 10:59:03] <elasticdog> ohadlevy: so I've narrowed it down...it works fine with straight puppetmasterd, but not when using passenger/httpd. when using apache, I get "hostname not match with the server certificate" in the logs | ||
| [2010/02/08 10:59:04] <ohadlevy> rmiller4pi8: and i didnt want yet another ca | ||
| [2010/02/08 10:59:23] <Volcane> kami: inside a define you must create unique resorces, if your defines require a single parent directory for exaple you must make those in a class | ||
| [2010/02/08 10:59:30] <elasticdog> is there a place to make sure apache is representing the hostname correctly? | ||
| [2010/02/08 10:59:32] <ohadlevy> elasticdog: try the host fqdn instead | ||
| [2010/02/08 10:59:33] @ yappy joined channel #puppet | ||
| [2010/02/08 10:59:36] <rmiller4pi8> ohadlevy: i have them sharing certs | ||
| [2010/02/08 11:00:02] <joe-mac> Pupeno: please read the wiki:TypeReference | ||
| [2010/02/08 11:00:06] <joe-mac> look under exec | ||
| [2010/02/08 11:00:31] @ Quit: jcape: Ping timeout: 256 seconds | ||
| [2010/02/08 11:00:41] <ohadlevy> rmiller4pi8: but then you need to generate another cert from the puppet cert AFAIK - which is not a big deal, the main issue that we tried running it on solaris for a while, it just didnt run | ||
| [2010/02/08 11:01:24] <yappy> I'm having issues fiding the clients facts from inside a report on the puppetmaster. Anyone have any pointers? | ||
| [2010/02/08 11:01:36] @ murkk joined channel #puppet | ||
| [2010/02/08 11:02:04] <ohadlevy> yappy: what are you trying to do? | ||
| [2010/02/08 11:02:46] <rmiller4pi8> ohadlevy: i'm literally using the same certs for both. but i've no doubt that outside the redhat family you are simply SOL with it. | ||
| [2010/02/08 11:02:51] <yappy> I'm after a dump of the clients facts to a file on check in. I can't see where I can get them from, from within the report | ||
| [2010/02/08 11:03:10] <yappy> Facter.facts gives me the puppetmasters - which I feel very silly for not catching earlier | ||
| [2010/02/08 11:03:20] <ohadlevy> yappy: its not part of the report | ||
| [2010/02/08 11:03:45] <kami> Volcane: I have _tried_ to pay attention to that and _think_ that all dirs which I use in the define itself, are non-singleton directories. But, I must be blind. | ||
| [2010/02/08 11:03:51] <ohadlevy> rmiller4pi8: interesting, so https://fedorahosted.org/func/wiki/FuncWithPuppet is simply wrong | ||
| [2010/02/08 11:03:58] @ eventi left channel #puppet () | ||
| [2010/02/08 11:04:14] <yappy> Ah, I thought the puppet client sent them over on check in, so I assumed you could access them from inside a report. bugger | ||
| [2010/02/08 11:04:48] <ohadlevy> yappy: what are you trying to achieve at the end? | ||
| [2010/02/08 11:04:48] <jbooth> rmiller4pi8: so Func looks interesting, but does it have anything to limit the parallelism? Like if I wanted to apply patches to say, 768 machines, but only have some number (say 32) patching at once to avoid overloading the patch server? | ||
| [2010/02/08 11:05:24] <yappy> A list of all the facts from a client in a text file. | ||
| [2010/02/08 11:05:35] <ohadlevy> yappy: no i imean after that :) | ||
| [2010/02/08 11:05:43] <rmiller4pi8> ohadlevy: what are you seeing that's wrong? it's Monday, so i could be missing something... | ||
| [2010/02/08 11:05:49] <kami> my oracle10g module (based on a recipe which is sited in the comment) is at http://pastie.org/814717 | ||
| [2010/02/08 11:05:58] <ohadlevy> yappy: for that just look in the vardir/yaml/facts | ||
| [2010/02/08 11:06:04] <rmiller4pi8> jbooth: if you call it as a python module rather than from CLI you can decide the parallelism at runtime | ||
| [2010/02/08 11:06:22] <kami> the directory $oracle10g::admindir is the one which is the problem | ||
| [2010/02/08 11:06:26] <Volcane> kami: what line number gives the problem? | ||
| [2010/02/08 11:06:35] <ohadlevy> rmiller4pi8: the whole part about using openssl to regenerate the certifictes | ||
| [2010/02/08 11:06:49] <kami> Volcane: 353 | ||
| [2010/02/08 11:06:54] <Volcane> kami: the trailing / suggests something else isnt being found like maybe $sidadmindir = "$admindir/$sid" if $sid is empty | ||
| [2010/02/08 11:07:05] <jbooth> rmiller4pi8: Hmm. How tightly can you control it? Just set to 32 and forget, or would is there a "okay run the next one" so I could actually have said script do load-limiting based on the server's suffering? | ||
| [2010/02/08 11:07:35] <yappy> ohadlevy: Ah, excellent - didn't know about that - thanks. | ||
| [2010/02/08 11:07:37] <kami> Volcane: OK. Will check that | ||
| [2010/02/08 11:07:49] <rmiller4pi8> jbooth: no, you can't really do that. you set the parallelism at module call time. i guess you could check for suffering and bail out, but then you might leave things in the middle. | ||
| [2010/02/08 11:08:10] <jbooth> Pity. Noted though, thanks... | ||
| [2010/02/08 11:08:15] @ Quit: ckauhaus: Ping timeout: 265 seconds | ||
| [2010/02/08 11:08:36] <ohadlevy> yappy: there are a few puppet reporting tools out there, they might solve your problem already :) | ||
| [2010/02/08 11:08:39] <rmiller4pi8> ohadlevy: the only openssl call on that page is removing the passphrase from the single CA private key | ||
| [2010/02/08 11:09:22] <kami> Volcane: I call it like this: http://pastie.org/814724 | ||
| [2010/02/08 11:09:47] <ohadlevy> rmiller4pi8: so you dont need to run it on the clients? hmm... maybe its too late here for me to read it properly :) | ||
| [2010/02/08 11:09:52] <Volcane> kami: show the actual error too pls | ||
| [2010/02/08 11:10:22] @ Quit: brothers: Read error: Connection reset by peer | ||
| [2010/02/08 11:10:44] * kami re-runs puppetd. | ||
| [2010/02/08 11:10:53] <rmiller4pi8> ohadlevy: correct | ||
| [2010/02/08 11:11:56] <ohadlevy> rmiller4pi8: oh well :) i guess its good to have many options :) | ||
| [2010/02/08 11:12:05] @ brothers joined channel #puppet | ||
| [2010/02/08 11:12:08] <kami> Volcane: http://pastie.org/814733 | ||
| [2010/02/08 11:12:37] @ Robbie__ joined channel #puppet | ||
| [2010/02/08 11:12:45] @ Quit: Jenza: Remote host closed the connection | ||
| [2010/02/08 11:13:05] <Volcane> kami: and does the line numbers in your paste align with the actual ones - so the error message shows a line and thats the same line number in your paste? | ||
| [2010/02/08 11:13:07] @ Quit: brothers: Read error: Connection reset by peer | ||
| [2010/02/08 11:13:12] @ brothers joined channel #puppet | ||
| [2010/02/08 11:13:26] @ Quit: Robbie_: Read error: Operation timed out | ||
| [2010/02/08 11:14:01] <ohadlevy> rmiller4pi8: i still think that thats only 20% of the problem, getting the right host list is usually the harder part | ||
| [2010/02/08 11:14:15] <kami> Volcane: Awfully sorry! No. | ||
| [2010/02/08 11:14:21] * kami counts lines | ||
| [2010/02/08 11:14:31] <Volcane> kami: heh | ||
| [2010/02/08 11:15:01] * kami is a moron. He knows. | ||
| [2010/02/08 11:15:58] <jb> odd | ||
| [2010/02/08 11:15:59] <jb> lots of daemons on my cent5.4 boxes randmoly restarted at 4:02AM yesterday | ||
| [2010/02/08 11:16:12] <ohadlevy> jb: logrotate ? | ||
| [2010/02/08 11:16:23] <kami> Volcane: 353 in the error message is 344 in my paste. | ||
| [2010/02/08 11:16:31] <jb> i don't think so | ||
| [2010/02/08 11:16:51] <gepetto> ::trac:: Whos Using Puppet edited @ http://reductivelabs.com/trac/puppet/wiki/WhosUsingPuppet?version=200 | ||
| [2010/02/08 11:17:20] @ jbartus left channel #puppet () | ||
| [2010/02/08 11:18:11] @ Quit: pacalm: Quit: pacalm | ||
| [2010/02/08 11:19:05] @ illsci left channel #puppet () | ||
| [2010/02/08 11:20:02] <Volcane> kami: so you do: $sidadmindir = $oracle10g::sidadmindir | ||
| [2010/02/08 11:20:10] <Volcane> kami: which will always yield the same directory | ||
| [2010/02/08 11:20:25] <Volcane> kami: then you make it many times with file in oracle10g::sid | ||
| [2010/02/08 11:22:54] <kami> Volcane: did I say I'm a moron? | ||
| [2010/02/08 11:23:16] <kami> Volcane: it was correct before I refactored something and then forgot about it. | ||
| [2010/02/08 11:23:22] <kami> Volcane: thank you! | ||
| [2010/02/08 11:23:24] <Volcane> kami: hehe ok | ||
| [2010/02/08 11:23:39] @ lohapuk is now known as lohapuk-away | ||
| [2010/02/08 11:24:01] * kami appreciates the kind help on #puppet | ||
| [2010/02/08 11:24:56] <Volcane> kami: you'll get help easiesr if you made 1 paste with code+error+node all in one go - its a lot less context switching to do to help someone - if it looks too hard to help someone i generally just dont | ||
| [2010/02/08 11:25:41] <kami> Volcane: thank you. And will try to remember that for next time. | ||
| [2010/02/08 11:25:47] <Volcane> kewl kewl | ||
| [2010/02/08 11:28:49] @ Quit: skiold: Quit: skiold | ||
| [2010/02/08 11:30:43] @ alfism joined channel #puppet | ||
| [2010/02/08 11:32:32] @ Quit: Ramonster: Quit: leaving | ||
| [2010/02/08 11:36:19] @ Quit: verwilst: Quit: Ex-Chat | ||
| [2010/02/08 11:38:31] @ ckauhaus joined channel #puppet | ||
| [2010/02/08 11:39:06] @ Quit: ptman: Ping timeout: 240 seconds | ||
| [2010/02/08 11:40:17] @ lak joined channel #puppet | ||
| [2010/02/08 11:43:03] @ Quit: LukeL_: Quit: Bye! | ||
| [2010/02/08 11:43:27] @ Quit: ckauhaus: Client Quit | ||
| [2010/02/08 11:43:30] @ jes5 joined channel #puppet | ||
| [2010/02/08 11:44:30] <kami> when I do $split_var = split ( $var, ':'), how do I access the first element of $split_var ? | ||
| [2010/02/08 11:47:38] @ Quit: lilmatt: Quit: lilmatt | ||
| [2010/02/08 11:51:07] @ sts left channel #puppet () | ||
| [2010/02/08 11:53:56] <fluxdude> I am running facter and am finding that it is giving me a ghost interface on network_eth0 when I am now only using eth1 | ||
| [2010/02/08 11:54:08] <fluxdude> eth0 was up a few minutes ago but I've intentionally shut it down | ||
| [2010/02/08 11:54:28] <fluxdude> and when I re-run facter it keeps giving the network details of network_eth0 as if it were a current fact | ||
| [2010/02/08 11:54:40] <elasticdog> ohadlevy: thanks for the help...got it all straightened out, has a typo with the rack config, so httpd wasn't loading it correctly | ||
| [2010/02/08 11:56:16] @ lilmatt joined channel #puppet | ||
| [2010/02/08 11:56:32] @ jes51 joined channel #puppet | ||
| [2010/02/08 11:56:34] @ Quit: jes5: Read error: Connection reset by peer | ||
| [2010/02/08 11:56:34] <fluxdude> how can I get facter to give me current facts only? | ||
| [2010/02/08 11:57:20] @ paxos joined channel #puppet | ||
| [2010/02/08 11:58:03] @ ptman joined channel #puppet | ||
| [2010/02/08 11:58:35] <Volcane> fluxdude: it should only give you current but i guess it might have a bug where it reports configured but down interfaces not sure | ||
| [2010/02/08 11:59:23] <Volcane> it runs ifconfig -a | ||
| [2010/02/08 11:59:35] <Volcane> so i guess if you just down'd it it will still show up | ||
| [2010/02/08 12:00:31] @ Quit: ptman: Read error: Operation timed out | ||
| [2010/02/08 12:03:04] @ ptman joined channel #puppet | ||
| [2010/02/08 12:03:46] @ benkevan joined channel #puppet | ||
| [2010/02/08 12:04:12] @ benkevan left channel #puppet () | ||
| [2010/02/08 12:06:40] <himynameisthor> woohoo | ||
| [2010/02/08 12:06:44] <himynameisthor> gettin my learn on | ||
| [2010/02/08 12:07:26] @ Quit: suso: Remote host closed the connection | ||
| [2010/02/08 12:08:03] @ Quit: shug: Quit: Leaving | ||
| [2010/02/08 12:08:20] @ Quit: bevans5446: Quit: bevans5446 | ||
| [2010/02/08 12:08:58] @ bevans5446 joined channel #puppet | ||
| [2010/02/08 12:10:08] @ Robbie_ joined channel #puppet | ||
| [2010/02/08 12:10:51] @ Quit: benlovell: Quit: benlovell | ||
| [2010/02/08 12:12:15] @ Quit: danielbln: Quit: Leaving. | ||
| [2010/02/08 12:13:53] @ Quit: Robbie__: Ping timeout: 258 seconds | ||
| [2010/02/08 12:15:38] @ Quit: lak: Quit: lak | ||
| [2010/02/08 12:15:57] @ lak joined channel #puppet | ||
| [2010/02/08 12:16:12] @ Quit: lak: Client Quit | ||
| [2010/02/08 12:17:37] @ Quit: ptman: Ping timeout: 240 seconds | ||
| [2010/02/08 12:18:20] @ Quit: mattwynne: Remote host closed the connection | ||
| [2010/02/08 12:18:52] <kami> Is there a way how I could split $name into different parts and use those parts in my definition? | ||
| [2010/02/08 12:19:46] <jrojas> does config_version apply per-environment? | ||
| [2010/02/08 12:20:05] <jrojas> kami: http://reductivelabs.com/trac/puppet/wiki/ReleaseNotes#id111 | ||
| [2010/02/08 12:21:00] @ shadoi joined channel #puppet | ||
| [2010/02/08 12:21:53] <kami> jrojas: thank you. But as a complete newbie, I don't know how to use it in a define. | ||
| [2010/02/08 12:22:06] <kami> I've tried several things but all give syntax errors. | ||
| [2010/02/08 12:22:29] <jrojas> http://docs.reductivelabs.com/references/stable/function.html | ||
| [2010/02/08 12:22:37] @ Quit: zazan: Ping timeout: 264 seconds | ||
| [2010/02/08 12:24:13] <kami> jrojas: thank you! The release notes reveal the problem: my puppet is too old. | ||
| [2010/02/08 12:25:00] @ lohapuk-away is now known as lohapuk | ||
| [2010/02/08 12:26:17] @ Quit: jaredrhine: Ping timeout: 240 seconds | ||
| [2010/02/08 12:26:43] @ tonyskapunk joined channel #puppet | ||
| [2010/02/08 12:28:10] @ Quit: moeian: Remote host closed the connection | ||
| [2010/02/08 12:28:21] @ Quit: tonyskapunk: Remote host closed the connection | ||
| [2010/02/08 12:29:07] @ tonyskapunk joined channel #puppet | ||
| [2010/02/08 12:29:31] @ ptman joined channel #puppet | ||
| [2010/02/08 12:29:41] @ Quit: fdz: Quit: Leaving. | ||
| [2010/02/08 12:32:28] @ Quit: poison: Remote host closed the connection | ||
| [2010/02/08 12:33:53] @ jcape joined channel #puppet | ||
| [2010/02/08 12:35:17] @ mgarfias left channel #puppet () | ||
| [2010/02/08 12:36:14] @ ckdake joined channel #puppet | ||
| [2010/02/08 12:36:30] @ bkevan joined channel #puppet | ||
| [2010/02/08 12:36:48] @ Quit: bkevan: Remote host closed the connection | ||
| [2010/02/08 12:39:04] <ckdake> is there a way to do negations in if/else statements in puppet? i.e. if (not $is_virtual) { } | ||
| [2010/02/08 12:39:30] <himynameisthor> if ! $is_virtual maybe? | ||
| [2010/02/08 12:39:31] <jrojas> wiki:LanguageTutorial | ||
| [2010/02/08 12:39:33] <gepetto> jrojas: wiki: wiki:LanguageTutorial is http://reductivelabs.com/trac/puppet/wiki/LanguageTutorial | ||
| [2010/02/08 12:39:40] <ckdake> the language tutorial seems to indicate that if (! $is_isvirtual) should work, but it doesnt seem to | ||
| [2010/02/08 12:40:05] @ benkevan joined channel #puppet | ||
| [2010/02/08 12:40:06] <ckdake> i.e. on the host, 'facter | grep is_virtual' says "true" but definitions inside of that if {} are not run | ||
| [2010/02/08 12:40:12] <jrojas> http://www.reductivelabs.com/trac/puppet/wiki/LanguageTutorial#id34 | ||
| [2010/02/08 12:40:20] <jrojas> http://www.reductivelabs.com/trac/puppet/wiki/LanguageTutorial#expressions | ||
| [2010/02/08 12:41:25] @ Quit: Robbie_: Remote host closed the connection | ||
| [2010/02/08 12:41:33] <ckdake> yeah, all those are already open in my browser :) | ||
| [2010/02/08 12:43:12] @ jes5 joined channel #puppet | ||
| [2010/02/08 12:43:12] @ Quit: jes51: Read error: Connection reset by peer | ||
| [2010/02/08 12:43:57] <benkevan> Ok.. I'm going to give puppet a shot.. cross your fingers for me (haha) | ||
| [2010/02/08 12:44:17] <himynameisthor> benkevan: no finger-crossing is necessary as long as you run --noop ;] | ||
| [2010/02/08 12:44:21] <ckdake> heres a simplified down example that shows this issue: http://pastebin.com/mec628fc | ||
| [2010/02/08 12:44:35] <ckdake> for a machine where "is_virtual" is false, ntp doesnt get installed | ||
| [2010/02/08 12:45:31] <benkevan> himynameisthor: --noop? I just installed it haha.. going to be quite a learning curve.. I already have a process in place to do some of the things puppet does.. but going to move from my home grown script (called updater, and stasher) and move to a more "standard" way of doing this (which I hope to achieve with puppet).. | ||
| [2010/02/08 12:46:20] <jrojas> ckdake: works for me | ||
| [2010/02/08 12:46:36] <Volcane> ckdake: what version puppet? | ||
| [2010/02/08 12:46:56] <ckdake> jrojas: hmm. 0.25.3-2.el5 | ||
| [2010/02/08 12:47:12] <jrojas> ckdake: are you sure you are getting a true/false value back from is_virtual? | ||
| [2010/02/08 12:47:27] <ckdake> thats what i see in the output of `facter` | ||
| [2010/02/08 12:47:29] <jrojas> ckdake: is ntp installed already? | ||
| [2010/02/08 12:47:30] <Volcane> maybe check for is_virtual == false or even "false" | ||
| [2010/02/08 12:47:52] <ckdake> jrojas: nope, removed it before running puppet on test box | ||
| [2010/02/08 12:47:59] <ckdake> Volcane: ill try that one | ||
| [2010/02/08 12:48:18] @ weaselkeeper joined channel #puppet | ||
| [2010/02/08 12:48:20] <Volcane> also just do notify{"is virtual: '${is_virtual}'": } | ||
| [2010/02/08 12:48:22] @ Quit: itguru: Ping timeout: 272 seconds | ||
| [2010/02/08 12:49:04] <ckdake> will that show up in puppet's log, or puppetmasters log? | ||
| [2010/02/08 12:49:19] <Volcane> on the client | ||
| [2010/02/08 12:49:23] <Volcane> run with --test | ||
| [2010/02/08 12:49:40] <ckdake> is virtual: 'false' | ||
| [2010/02/08 12:49:46] <ckdake> so checking string equality may be the way to go | ||
| [2010/02/08 12:49:50] <Volcane> k | ||
| [2010/02/08 12:50:35] <ckdake> yep: `if $is_virtual == 'false' {` | ||
| [2010/02/08 12:51:22] <ckdake> that works! in earlier versions of puppet, `if ( $is_virtual) {} else {` seemed to work, but it doesn't seem to now. thankfully there are only two things to change :) | ||
| [2010/02/08 12:51:24] <ckdake> thanks for your help | ||
| [2010/02/08 12:53:19] <gepetto> ::trac:: Release Notes edited @ http://reductivelabs.com/trac/puppet/wiki/ReleaseNotes?version=58 | ||
| [2010/02/08 12:53:49] <benkevan> damn i have alot to learn with this | ||
| [2010/02/08 12:54:06] @ poison joined channel #puppet | ||
| [2010/02/08 12:57:12] @ glarizza joined channel #puppet | ||
| [2010/02/08 12:58:20] @ suchu joined channel #puppet | ||
| [2010/02/08 12:58:59] <glarizza> Hey guys, stupid question: If you have file {"/users/students/library/preferences.com.apple.desktop.plist": ensure => absent} will it actually remove the file if it finds it, or does 'ensure => absent' not work that way? | ||
| [2010/02/08 12:59:26] <rgsteele> Just out of curiosity, is the limit of Puppet's LDAP capacities limited to fetching node configurations? I was wondering if I could get some finer granularity, such as classes which can fetch user information or daemon configurations from the directory. | ||
| [2010/02/08 12:59:57] <ckdake> glarizza: this explains it pretty well: http://docs.reductivelabs.com/references/stable/type.html#file | ||
| [2010/02/08 13:01:13] <rgsteele> Hm, actually, the puppetclass LDAP attribute can probably do what I want. | ||
| [2010/02/08 13:01:26] <glarizza> That's what I thought - I must have something buggered. Thanks. | ||
| [2010/02/08 13:02:03] @ Quit: TREllis: Quit: leaving | ||
| [2010/02/08 13:02:30] @ Quit: benkevan: Remote host closed the connection | ||
| [2010/02/08 13:02:35] * ckdake watches "snmp trap alertSystemUp"s come in from lots of machines as the `$is_virtual == 'false` is rolled out | ||
| [2010/02/08 13:03:57] @ kami` joined channel #puppet | ||
| [2010/02/08 13:05:45] @ Quit: kami: Ping timeout: 256 seconds | ||
| [2010/02/08 13:05:51] @ autowidget joined channel #puppet | ||
| [2010/02/08 13:06:21] <whack> cd | ||
| [2010/02/08 13:08:24] @ Quit: glarizza: Quit: glarizza | ||
| [2010/02/08 13:09:34] <weaselkeeper> Just installed puppet on CentOS 5.4, fresh clean install, following the simple puppet start on Reductive Labs docs, getting the following error. | ||
| [2010/02/08 13:09:35] <weaselkeeper> [root@rx9000 classes]# /etc/init.d/puppetmaster start | ||
| [2010/02/08 13:09:35] <weaselkeeper> Starting puppetmaster: /usr/sbin/puppetmasterd:83:in `require': no such file to load -- puppet (LoadError) | ||
| [2010/02/08 13:09:35] <weaselkeeper> from /usr/sbin/puppetmasterd:83 | ||
| [2010/02/08 13:09:35] <weaselkeeper> No oddball packages installed, the puppet and puppet-server packages came from EPEL, Version 0.24.8-4, anyone using puppetmaster from CentOS 5 ? | ||
| [2010/02/08 13:09:39] @ Quit: suchu: Ping timeout: 276 seconds | ||
| [2010/02/08 13:09:58] <jrojas> weaselkeeper: use pastie.org | ||
| [2010/02/08 13:10:13] <weaselkeeper> sure thing. | ||
| [2010/02/08 13:10:39] <jrojas> weaselkeeper: sounds like you donthave rubygems or you have a weird ruby that is in a diff place than the puppet rpm you installed | ||
| [2010/02/08 13:10:51] <fluxdude> Volcane: you were exactly right by the way about facter's behaviour | ||
| [2010/02/08 13:11:00] <fluxdude> regarding ifconfig -a | ||
| [2010/02/08 13:11:03] <fluxdude> tnx | ||
| [2010/02/08 13:11:05] <fluxdude> thx | ||
| [2010/02/08 13:11:30] @ svend joined channel #puppet | ||
| [2010/02/08 13:11:42] <weaselkeeper> ruby, rubygems are from centos default, puppet, puppet-server, from Fedora's epel. | ||
| [2010/02/08 13:12:29] <weaselkeeper> ah, but I think you have hit the nail. | ||
| [2010/02/08 13:12:45] <weaselkeeper> This is a 64bit system, and RH is retarded on 64 bit library locations | ||
| [2010/02/08 13:13:02] @ danielbln joined channel #puppet | ||
| [2010/02/08 13:13:11] <weaselkeeper> rubygems, and ruby are in /usr/lib64, whereas puppet, is in /usr/lib. That is annoying | ||
| [2010/02/08 13:13:14] @ dan__t joined channel #puppet | ||
| [2010/02/08 13:14:53] @ gpled joined channel #puppet | ||
| [2010/02/08 13:15:23] <joe-mac> yea, the joys of rubygems | ||
| [2010/02/08 13:15:48] <weaselkeeper> don't get me started. | ||
| [2010/02/08 13:15:55] <weaselkeeper> Puppet I like, ruby... not so much | ||
| [2010/02/08 13:16:26] <z00dax> rubygems that are not really archspecific should be in /usr/lib/ | ||
| [2010/02/08 13:16:40] @ jaredrhine joined channel #puppet | ||
| [2010/02/08 13:16:51] @ Quit: themurph: Quit: themurph | ||
| [2010/02/08 13:17:00] <weaselkeeper> in which case, ruby should be able to see them, as well as the arch specific ones, yes? | ||
| [2010/02/08 13:17:13] <Volcane> do you have /usr/lib/ruby/site_ruby/1.8 ? | ||
| [2010/02/08 13:17:17] <Volcane> puppet should be in there | ||
| [2010/02/08 13:17:51] <Volcane> if not, you got a weird puppet package - lime maybe one for rhel4 | ||
| [2010/02/08 13:17:53] <weaselkeeper> yes, that was installed via the rpm from epel for puppet/puppet-server | ||
| [2010/02/08 13:17:55] <gpled> do most yum puppeteers, host their own repositories? | ||
| [2010/02/08 13:18:02] @ Quit: jaredrhine: Client Quit | ||
| [2010/02/08 13:18:28] <z00dax> weaselkeeper: there is no rubygems in centos | ||
| [2010/02/08 13:18:46] <ckdake> gpled: i use EPEL, and that seems to be pretty common | ||
| [2010/02/08 13:18:49] <z00dax> so i am guessing you got that from epel as well ? which would, yes - build all noarch gems into /usr/lib/ | ||
| [2010/02/08 13:19:07] <z00dax> ckdake: I think it has a lot to do with site size | ||
| [2010/02/08 13:19:23] <weaselkeeper> you're right, I think that came from epel | ||
| [2010/02/08 13:19:31] <jrojas> Xen VM + Hudson + Mock + SCM == automagic rpm building for my platforms :) | ||
| [2010/02/08 13:20:00] <weaselkeeper> so is there a way to tell ruby to look both in /usr/lib and /usr/lib64 paths for stuff? | ||
| [2010/02/08 13:20:18] <dan__t> hi. | ||
| [2010/02/08 13:21:06] <geekguy> so the version of puppet in openbsd ports is 0.24.4p1, should I just install .25.4 from source instead? | ||
| [2010/02/08 13:21:40] <Volcane> weaselkeeper: what ruby stuff do you have in lib64 thats puppet related? that should all be in /usr/lib/ruby/site_ruby | ||
| [2010/02/08 13:22:44] <himynameisthor> who's here at SFSU with me? | ||
| [2010/02/08 13:22:53] <himynameisthor> cough loudly | ||
| [2010/02/08 13:23:14] @ Quit: giskard: Remote host closed the connection | ||
| [2010/02/08 13:23:46] <weaselkeeper> vollmer: none that's puppet related that I can see, but ruby itself is. In perl I can add stuff to the INC, is there an equiv for ruby? | ||
| [2010/02/08 13:24:29] <jrojas> weaselkeeper: its stored in the 4: variable i think | ||
| [2010/02/08 13:24:31] <jrojas> err $: | ||
| [2010/02/08 13:24:42] <jrojas> ie: $:.push("/path") | ||
| [2010/02/08 13:24:43] <vollmer> I've been mis-pinged | ||
| [2010/02/08 13:24:55] <jrojas> or is it $LOAD_PATH ? | ||
| [2010/02/08 13:24:55] <vollmer> Volcane: | ||
| [2010/02/08 13:24:56] @ lak joined channel #puppet | ||
| [2010/02/08 13:25:32] <jrojas> weaselkeeper:in irb do: puts$:.sort | ||
| [2010/02/08 13:25:42] <jrojas> that will tell you where it is looking | ||
| [2010/02/08 13:26:35] <weaselkeeper> all in /usr/lib64, ruby neophyte here, what's the 4: variable? | ||
| [2010/02/08 13:26:43] <jrojas> $: | ||
| [2010/02/08 13:26:46] <jrojas> not 4: | ||
| [2010/02/08 13:26:55] <weaselkeeper> k | ||
| [2010/02/08 13:26:56] <jrojas> $LOAD_PATH == $: | ||
| [2010/02/08 13:28:25] <yappy> I'm having problems getting puppetmaster to pick up my new type - does http://pastie.org/814958 look like a sensible tree? | ||
| [2010/02/08 13:29:20] <gpled> if the client has more then one repository, is there a way to get puppet to tell the client, get this yum package from this repository, and get that package, from that repository? | ||
| [2010/02/08 13:29:53] <Volcane> gpled: yeah, configure yum properly | ||
| [2010/02/08 13:29:57] <weaselkeeper> ok, $:.push("/usr/lib/ruby/site_ruby/1.8") | ||
| [2010/02/08 13:29:57] <weaselkeeper> clears up the problem in that instance, is there a way to do that globally? either in the startup script for puppet or similar? | ||
| [2010/02/08 13:30:12] <gpled> Volcane: on the clients? | ||
| [2010/02/08 13:30:20] <Volcane> weaselkeeper: if you need to do this on a centos machine, you broke it very badly and should just start fresh | ||
| [2010/02/08 13:30:23] <weaselkeeper> gpled: by default, yum will grab the latest overall, from all the repos | ||
| [2010/02/08 13:30:41] <Volcane> gpled: yes, yum has a config file and repos can be set what packages to get from where - its not something you can tell puppet to just arbitrarily do | ||
| [2010/02/08 13:30:56] <Volcane> gpled: puppet will just run yum install with optionally a version | ||
| [2010/02/08 13:31:43] <tim|imac> hey all, I can't find in the language tutorial any reference to an else-if construction... is there any? or should I just use several if's without else? | ||
| [2010/02/08 13:31:47] <weaselkeeper> Volcane: you say that /usr/lib should be used for arch independant gems/packages, how des ruby handle a mix of arch and noarch packages? | ||
| [2010/02/08 13:32:08] <Volcane> weaselkeeper: it just does the right thing - arch specific ones goes to lib64 and non arch one goes into lib | ||
| [2010/02/08 13:32:11] @ Quit: dotNox: Quit: Lost terminal | ||
| [2010/02/08 13:32:28] @ ambroff joined channel #puppet | ||
| [2010/02/08 13:32:35] <weaselkeeper> so where is the search path defined? | ||
| [2010/02/08 13:32:49] @ Quit: kolla: Remote host closed the connection | ||
| [2010/02/08 13:35:03] <Volcane> convinced you installed the wrong package from the wrong source on the wrong centos or something very weird | ||
| [2010/02/08 13:35:38] <jrojas> weaselkeeper: on my default centos 5.2 install, the $: var includes /usr/lib/ruby/1.8 and /usr/lib/ruby/site_ruby | ||
| [2010/02/08 13:36:02] <jrojas> verified the x86_64 package of puppet in epel installs in /usr/lib/ruby/site_ruby/1.8 | ||
| [2010/02/08 13:36:57] @ Quit: jes5: Ping timeout: 240 seconds | ||
| [2010/02/08 13:36:57] <weaselkeeper> jrojas: that's where it's installed on my box too. What sets the $: var ? | ||
| [2010/02/08 13:37:08] <jrojas> weaselkeeper: umm...ruby | ||
| [2010/02/08 13:37:16] <jrojas> what version of ruby? | ||
| [2010/02/08 13:40:35] <weaselkeeper> 1.8.6 atm, but we built it, so it's likely an issue with that. Just trying to figure out how ruby knows where to search for stuff. | ||
| [2010/02/08 13:40:43] @ re_ joined channel #puppet | ||
| [2010/02/08 13:40:53] <Volcane> holy hell, you neglected till now to mention you built your own | ||
| [2010/02/08 13:40:58] <Volcane> insisting its just centos ruby? | ||
| [2010/02/08 13:40:59] <jrojas> weaselkeeper: there is your problem. not saying you dont know how to build ruby, but yeah thats the issue | ||
| [2010/02/08 13:41:17] <weaselkeeper> good, cause I didn't build this one, predates me here :P | ||
| [2010/02/08 13:41:30] <Volcane> "ruby, rubygems are from centos default" | ||
| [2010/02/08 13:41:52] <weaselkeeper> yep, that's an error on my part, it had the word centos in it, but it's built here | ||
| [2010/02/08 13:41:55] <jrojas> head meet desk. desk meet head. | ||
| [2010/02/08 13:42:41] <weaselkeeper> and as for ruby, yeah, I don't know much ruby, managed to avoid it so far, new job, new challenges, oh joy. | ||
| [2010/02/08 13:43:07] <joe-mac> all this magical AD changing DNS bullshit aggravates the fuck out of me | ||
| [2010/02/08 13:43:57] <weaselkeeper> which leads me back to the question of how does ruby know what path to search for gems/packages? Obviously, If I can change that, looks like my problem is solved. | ||
| [2010/02/08 13:44:16] <Volcane> weaselkeeper: pretty sure its compile time | ||
| [2010/02/08 13:44:21] <weaselkeeper> joe-mac: yeah, AD DNS is a joy to play with, it's fun, like poking a red hot poker in your eye is fun. | ||
| [2010/02/08 13:44:27] <joe-mac> no no best part is | ||
| [2010/02/08 13:44:36] <joe-mac> someone thought it would be cool to have a random bind server running elsewehre | ||
| [2010/02/08 13:44:38] <weaselkeeper> Volcane: that's ... awkward. | ||
| [2010/02/08 13:44:39] <joe-mac> justf or fun | ||
| [2010/02/08 13:45:02] <joe-mac> i need a vacation so bad | ||
| [2010/02/08 13:45:13] <weaselkeeper> joe-mac: yeah, rule #1 about AD-DNS is "thou shalt have no other DNS servers than mine!" | ||
| [2010/02/08 13:45:19] <gepetto> weaselkeeper: joe-mac: #1 is http://projects.reductivelabs.com/issues/show/1 "Puppet - Feature #1: Differentiate classes from definitions - ReductiveLabs.com" | ||
| [2010/02/08 13:45:48] <joe-mac> yea, which i honestly wouldn't mind weaselkeeper | ||
| [2010/02/08 13:46:01] <joe-mac> the fact is i have to jump through hoops to add a freakin a record because of this | ||
| [2010/02/08 13:46:46] <re_> For each server in my cluster I need to add a unique static ip on eth1 to /etc/network/interfaces (i'm running debian) without messing up anything set on eth0. I'm wondering what would be best way to automate this with puppet? Have an exec that concats the eth1 bits with an unless => "cat /etc/network/interfaces | grep eth1" ?? | ||
| [2010/02/08 13:47:10] <Volcane> weaselkeeper: your best option is just to take the puppet srcrpm and build it on this broken machine of yours - you should get a rpm thats broken in a compatible way | ||
| [2010/02/08 13:47:14] <joe-mac> re_: i have a define to do all that but it is a lot of files and i am too busy to paste it | ||
| [2010/02/08 13:47:34] <Volcane> weaselkeeper: but anything ruby related you download from epel/redhat/centos/fedora just wont work for you | ||
| [2010/02/08 13:48:10] <re_> joe-mac: lol =p is it available online by any chance? | ||
| [2010/02/08 13:48:27] @ jes5 joined channel #puppet | ||
| [2010/02/08 13:49:30] @ Quit: svend: Remote host closed the connection | ||
| [2010/02/08 13:49:32] <weaselkeeper> yeah, rebuilting puppet is not the answer, rebuilding our ruby is, but, *how* do I change the ruby rpm to search in both locations for gems/packages? that's the question. If ruby needs to search in both locations for arch specific and not specific packages, how do it know? | ||
| [2010/02/08 13:50:13] <Volcane> look how the centos ruby rpms are built | ||
| [2010/02/08 13:50:48] <jrojas> weaselkeeper: is there a sepcific reason you needed ruby rebuilt? | ||
| [2010/02/08 13:50:55] <jrojas> cant you just use the default ruby packages? | ||
| [2010/02/08 13:52:46] <weaselkeeper> at the time, ruby wasn't available for our platform, so we built it, and have continued to carry it forward (like 2 years ago) we have some 3000 installs, so it's not a simple task to change that. | ||
| [2010/02/08 13:53:51] <jrojas> dssh rpm -qva|grep -i ruby| sudo xargs rpm -e --nodeps | ||
| [2010/02/08 13:53:52] <jrojas> :) | ||
| [2010/02/08 13:54:19] <jrojas> or you edit the spec file from the default src.rpm of ruby for centos, and add Obsoletes: your.old.package | ||
| [2010/02/08 13:54:28] @ ckdake left channel #puppet () | ||
| [2010/02/08 13:55:40] <eric0> if you have control over the startup script you can just set RUBYLIB in the environment | ||
| [2010/02/08 13:56:02] <weaselkeeper> and it looks like CentOS 5.4 still doesn't carry ruby 1.8.6, sigh | ||
| [2010/02/08 13:56:11] <eric0> i'm doing that now, it's not aesthetically pleasing but works fine. | ||
| [2010/02/08 13:56:15] <weaselkeeper> eric0: thanks, that will work for me. | ||
| [2010/02/08 13:56:29] <weaselkeeper> I can rebuild the package with that change easy enough | ||
| [2010/02/08 13:57:43] <weaselkeeper> jrojas: ok, not a simple task to change it in such a way as I don't have to look for work afterwards "P | ||
| [2010/02/08 13:58:53] <jrojas> weaselkeeper: providing future platform stability to avoid unecessary hacks makes you loose your job? | ||
| [2010/02/08 14:00:05] <weaselkeeper> killing a few hundred or a thousand custom customer rails apps by removing ruby from underneath them, would likely have an impact on perceived job performance. | ||
| [2010/02/08 14:00:09] @ Quit: labrown: Quit: Leaving | ||
| [2010/02/08 14:00:27] <Volcane> weaselkeeper: why would it? thats by design | ||
| [2010/02/08 14:00:53] <weaselkeeper> doing an rpm -e --nodeps ruby would kind of do that. | ||
| [2010/02/08 14:00:55] @ giskard joined channel #puppet | ||
| [2010/02/08 14:00:58] @ Quit: lak: Quit: lak | ||
| [2010/02/08 14:01:31] @ lak joined channel #puppet | ||
| [2010/02/08 14:03:04] @ Quit: giskard: Read error: Connection reset by peer | ||
| [2010/02/08 14:04:32] <jb> we have a AD/BIND hybrid environment | ||
| [2010/02/08 14:04:33] <jb> :/ | ||
| [2010/02/08 14:05:29] <weaselkeeper> jb: no doubt your joy is unbound :P | ||
| [2010/02/08 14:08:38] @ themurph joined channel #puppet | ||
| [2010/02/08 14:08:53] @ jaredrhine joined channel #puppet | ||
| [2010/02/08 14:09:44] @ Quit: themurph: Client Quit | ||
| [2010/02/08 14:09:59] @ Quit: Pupeno: Quit: Pupeno | ||
| [2010/02/08 14:12:12] <whack> AD actually works ok with bind | ||
| [2010/02/08 14:12:27] <whack> though the lat time I used it, AD wasn't capable of doing DNSSEC, so you had to whitelist nsupdates by IP. | ||
| [2010/02/08 14:12:38] @ themurph joined channel #puppet | ||
| [2010/02/08 14:13:34] <jb> weaselkeeper: actually, it works great :) | ||
| [2010/02/08 14:13:47] <jb> AD is only authoritative for the integrated AD dynamic zone | ||
| [2010/02/08 14:13:52] <jb> it forwards everything else to BIND | ||
| [2010/02/08 14:14:02] <jb> but, a bit OT :) | ||
| [2010/02/08 14:14:20] @ Quit: bobbyz: Ping timeout: 265 seconds | ||
| [2010/02/08 14:14:43] @ Quit: geekguy: Quit: Page closed | ||
| [2010/02/08 14:18:48] @ ashp joined channel #puppet | ||
| [2010/02/08 14:18:49] @ Quit: ashp: Client Quit | ||
| [2010/02/08 14:22:30] <duritong> oh github uses puppet | ||
| [2010/02/08 14:22:57] <jrojas> duritong: yes they kinda revealed that in their outage statement | ||
| [2010/02/08 14:22:58] <jrojas> :P | ||
| [2010/02/08 14:25:40] <gpled> can you have puppet add a yum repository to a client? | ||
| [2010/02/08 14:26:14] <jrojas> gpled: look at the wiki:TypeReference | ||
| [2010/02/08 14:26:22] <jrojas> there is a yumrepo type | ||
| [2010/02/08 14:28:36] <duritong> jrojas: yeah read it right now. a comment mention that they used to use chef. but I haven't been aware of that either | ||
| [2010/02/08 14:31:07] <kami`> Hello, after installing the current puppet, I still cannot use split correctly: http://pastie.org/815067 | ||
| [2010/02/08 14:35:06] @ Quit: danielbln: Quit: Leaving. | ||
| [2010/02/08 14:37:35] <Volcane> kami`: you cant access array members in puppet like that | ||
| [2010/02/08 14:37:54] <Volcane> only in templates | ||
| [2010/02/08 14:38:00] <jrojas> kami`: read this: http://groups.google.com/group/puppet-users/browse_thread/thread/b413277cd3e08787 | ||
| [2010/02/08 14:53:25] @ mrpots joined channel #puppet | ||
| [2010/02/08 14:53:59] @ Quit: jaredrhine: Ping timeout: 256 seconds | ||
| [2010/02/08 14:54:33] @ ezmobius joined channel #puppet | ||
| [2010/02/08 14:59:14] @ Ramonster joined channel #puppet | ||
| [2010/02/08 15:00:55] <tessier> I have a few users that I want to create everywhere. I could set up separate user and group types but it seems like it might be better to put them in an array. But then I'm not sure if I can say things like home => '/home/$userlist', and I'm not sure how I could fill in a comment => 'Real Name' without having some sort of two dimensional array or something. | ||
| [2010/02/08 15:02:19] <tessier> Are puppet configs/arrays really just ruby? If so I bet there is a way to do a two dimensional array... but not being familiar with ruby I can't say if it is or not. | ||
| [2010/02/08 15:05:38] <jrojas> tessier: not really, on the ruby part. and on the two dimensional arrays (hashes?) im not sure, but i probably could type faster than it would take to figure it out :) | ||
| [2010/02/08 15:05:51] @ xcrracer_ joined channel #puppet | ||
| [2010/02/08 15:06:06] <tessier> jrojas: Ok, just wondering. I'll just copy/paste it all out then. | ||
| [2010/02/08 15:06:55] @ ckauhaus joined channel #puppet | ||
| [2010/02/08 15:07:16] <whack> you probably want a custom define | ||
| [2010/02/08 15:07:32] <whack> so you can just say myuser { ["user1", "user2", ...]: ensure => present } | ||
| [2010/02/08 15:07:49] <whack> and 'myuser' would create the user resource with the proper name/homedir/etc | ||
| [2010/02/08 15:07:53] @ Quit: toi: Ping timeout: 260 seconds | ||
| [2010/02/08 15:08:07] <jrojas> whack: sounds like a lot of work for something that is easily managed already | ||
| [2010/02/08 15:08:10] <tessier> I am having a harder time getting my head around puppet than I did cfengine. | ||
| [2010/02/08 15:08:30] @ Quit: themurph: Quit: themurph | ||
| [2010/02/08 15:09:15] @ Quit: jcape: Remote host closed the connection | ||
| [2010/02/08 15:09:30] @ jcape joined channel #puppet | ||
| [2010/02/08 15:11:56] @ jaredrhine joined channel #puppet | ||
| [2010/02/08 15:14:41] <tessier> In puppet do you end up defining every single machine in nodes.pp or do you use some fact or other conditional to determine what should be done to a machine? | ||
| [2010/02/08 15:15:42] <fsweetser> for small/simple setups, you define them all in nodes.pp | ||
| [2010/02/08 15:15:56] <fsweetser> for slightly more complex ones, you can use regular expression matching nodes in site.pp | ||
| [2010/02/08 15:16:05] <tessier> ah | ||
| [2010/02/08 15:16:08] @ themurph joined channel #puppet | ||
| [2010/02/08 15:16:10] <jrojas> tessier: you can match regex in nodes.pp afaik | ||
| [2010/02/08 15:16:15] <jrojas> err | ||
| [2010/02/08 15:16:18] <jrojas> too slow | ||
| [2010/02/08 15:16:19] <fsweetser> for larger setups, you use an external script that you define | ||
| [2010/02/08 15:16:27] <Volcane> tessier: or use theforeman.org or use wiki:ExternalNodes | ||
| [2010/02/08 15:16:32] <gepetto> Volcane: tessier: wiki:ExternalNodes is http://reductivelabs.com/trac/puppet/wiki/ExternalNodes | ||
| [2010/02/08 15:16:38] <fsweetser> puppetmaster calls the script, and you implement whatever business logic you like to return classes for a given node | ||
| [2010/02/08 15:17:08] <fsweetser> yep, theforeman is an excellent example of an external nodes script | ||
| [2010/02/08 15:17:31] <tessier> Ok. So have a bunch of database machines in my db cluster. So I have a basenode, I will make a dbs which inherits basenode and will make each of my individual db machine definitions inherit dbs. So I only have to specify the special classes for my dbs once. | ||
| [2010/02/08 15:18:01] <Volcane> easier to just make a class db_node | ||
| [2010/02/08 15:18:10] <Volcane> put your stuff there and do include db_node whenever u need a db | ||
| [2010/02/08 15:18:33] @ Quit: robinbowes: Quit: Leaving | ||
| [2010/02/08 15:18:56] @ Quit: bug: Quit: bug | ||
| [2010/02/08 15:19:07] <tessier> Volcane: Then I would do include db_node for each db node I define? Seems like six of one and half dozen of another. | ||
| [2010/02/08 15:19:49] <Volcane> tessier: yes, but inheritance has a lot of shortcoming/weirdness and general suckyness | ||
| [2010/02/08 15:19:57] <Volcane> see the FAQ about that | ||
| [2010/02/08 15:20:03] <tessier> Ah. Well that is unfortunate. Ok, thanks. | ||
| [2010/02/08 15:20:42] @ Quit: nevyn: Ping timeout: 252 seconds | ||
| [2010/02/08 15:21:51] @ Quit: ambroff: Ping timeout: 245 seconds | ||
| [2010/02/08 15:22:33] @ nevyn joined channel #puppet | ||
| [2010/02/08 15:23:04] @ Quit: unxfrek: Quit: Leaving | ||
| [2010/02/08 15:23:15] @ robinbowes joined channel #puppet | ||
| [2010/02/08 15:25:34] @ gebi joined channel #puppet | ||
| [2010/02/08 15:26:45] @ Quit: nevyn: Ping timeout: 252 seconds | ||
| [2010/02/08 15:27:29] @ mvn071 joined channel #puppet | ||
| [2010/02/08 15:31:56] @ giskard joined channel #puppet | ||
| [2010/02/08 15:33:16] @ bobbyz joined channel #puppet | ||
| [2010/02/08 15:33:59] @ nevyn joined channel #puppet | ||
| [2010/02/08 15:34:11] @ benkevan joined channel #puppet | ||
| [2010/02/08 15:35:00] @ verwilst joined channel #puppet | ||
| [2010/02/08 15:35:15] @ gpled left channel #puppet ("Ex-Chat") | ||
| [2010/02/08 15:36:05] <dan__t> Damnit. | ||
| [2010/02/08 15:36:07] <dan__t> Got d/c'd. | ||
| [2010/02/08 15:36:28] <dan__t> Don't suppose you guys got what I had posted a few mins ago then about templating exceptions based on an applied class? | ||
| [2010/02/08 15:36:50] <tessier> dan__t: Nope | ||
| [2010/02/08 15:37:16] <dan__t> I want to take a template, erb, and apply a given line of a given class is applied | ||
| [2010/02/08 15:37:31] <dan__t> so like if class::subclass::instance is applied, then this line will be included | ||
| [2010/02/08 15:38:06] @ ccaum joined channel #puppet | ||
| [2010/02/08 15:38:51] <himynameisthor> dan__t: you i don't know if class names are passed through to the erb, but you could define a variable inside the class instance and check for its existance inside the erb | ||
| [2010/02/08 15:38:55] <tim|imac> how do i disable checksum checking for a file{} resource again? | ||
| [2010/02/08 15:38:57] <tim|imac> nosum? | ||
| [2010/02/08 15:39:23] <joe-mac> check the checksum parameter in the type reference for file | ||
| [2010/02/08 15:39:24] <joe-mac> i forget | ||
| [2010/02/08 15:39:35] <tim|imac> yeah, it's not in there :| | ||
| [2010/02/08 15:39:39] <ccaum> Hey everyone. I have a class with a service type that has ensure => running as a parameter. I need a subclass to remove the ensure parameter all together, not overwrite it. Is there a way to do that? | ||
| [2010/02/08 15:39:48] <tim|imac> http://docs.reductivelabs.com/references/stable/type.html#file | ||
| [2010/02/08 15:39:59] <tim|imac> although it hints that it's possible... | ||
| [2010/02/08 15:40:02] <dan__t> I sure could, yeah. Just wanted to make it simple, himynameisthor | ||
| [2010/02/08 15:40:14] <joe-mac> glahhh i am so tird | ||
| [2010/02/08 15:40:19] <joe-mac> brb moar coffee | ||
| [2010/02/08 15:40:23] <dan__t> I've done this before with setting variables inside of the class | ||
| [2010/02/08 15:41:05] <joe-mac> can somebody tell me where you configure repo sources for gems? | ||
| [2010/02/08 15:41:26] <ccaum> I tried just seeing if the subclass was defined and make the service resource accordingly, but that didn't work | ||
| [2010/02/08 15:43:13] @ weaselkeeper left channel #puppet () | ||
| [2010/02/08 15:43:15] @ bug joined channel #puppet | ||
| [2010/02/08 15:43:45] @ warreng joined channel #puppet | ||
| [2010/02/08 15:43:51] <ccaum> I've got this so far: http://pastie.org/815198. So if I include the monit::ha class in another class, it should use the service resource without the ensure parameter | ||
| [2010/02/08 15:43:55] <ccaum> But it no workie | ||
| [2010/02/08 15:49:13] @ Quit: bug: Read error: Connection reset by peer | ||
| [2010/02/08 15:49:36] @ Quit: bobbyz: Ping timeout: 240 seconds | ||
| [2010/02/08 15:49:43] @ bug joined channel #puppet | ||
| [2010/02/08 15:51:13] <benkevan> does puppet support zypper repo's? Or only yum? (sorry totally new to puppet.. and am actually only on the configuration of the puppetmaster | ||
| [2010/02/08 15:54:03] <ccaum> benkevan: I don't see zypper in the list: http://docs.reductivelabs.com/references/stable/type.html#package | ||
| [2010/02/08 15:54:07] <ccaum> But I don't know what zypper is | ||
| [2010/02/08 15:54:34] <elasticdog> I had the exact same issue as this guy: http://is.gd/7XLtI and solved it by re-adding the "RequestHeader" lines from the Passenger wiki page | ||
| [2010/02/08 15:54:34] <jrojas> benkevan: #1223 maybe? | ||
| [2010/02/08 15:54:36] <gepetto> jrojas: benkevan: #1223 is http://projects.reductivelabs.com/issues/show/1223 "Puppet - Bug #1223: rug not on OpenSuSE 10.3 anymore - ReductiveLabs.com" | ||
| [2010/02/08 15:54:36] <ccaum> benkevan: Oh sorry, you said repos. | ||
| [2010/02/08 15:56:20] <benkevan> jrojas: Thank you.. It seems as though the target version for a zypper provider is .25.5 .. I'll have to look to see what their release roadmap is.. | ||
| [2010/02/08 15:56:23] <jrojas> elasticdog: the X-SSL and X-Client.. stuff? | ||
| [2010/02/08 15:56:33] <benkevan> and that's Due in 20 days (02/28/2010) .. :) | ||
| [2010/02/08 15:56:50] <jrojas> benkevan: if someone is working on it there must be a need, which means you may want to search for a feature request if not open one regarding the matter | ||
| [2010/02/08 15:58:06] <elasticdog> jrojas: yep | ||
| [2010/02/08 15:58:18] <dan__t> Just thinking aloud here... going to have puppet manage some dev environments. I basically need to edit .bashrc and drop another few files in a user's home dir on a per-developer basis. Anyone have much experience doing something like this? Not quite sure on how to go about doing this. | ||
| [2010/02/08 15:59:17] <jrojas> define the homedirs as a file, have the user's require the home dirs, point the file definition's source to /etc/skel recurse=true etc.. .. ? | ||
| [2010/02/08 15:59:41] <dan__t> Yeah just didn't want to have to specify user home dirs | ||
| [2010/02/08 15:59:48] <jrojas> hm.. | ||
| [2010/02/08 15:59:51] <dan__t> Maybe I can loop for dirs in /home | ||
| [2010/02/08 15:59:55] <jrojas> nah | ||
| [2010/02/08 15:59:58] <jrojas> for loops are boring | ||
| [2010/02/08 16:00:02] <jrojas> virtual resources? | ||
| [2010/02/08 16:00:05] @ lohapuk is now known as lohapuk-awak | ||
| [2010/02/08 16:00:07] <dan__t> Perhaps... | ||
| [2010/02/08 16:00:09] @ lohapuk-awak is now known as lohapuk-away | ||
| [2010/02/08 16:00:39] <dan__t> Just don't want to have to manually APPLY this for every user manually | ||
| [2010/02/08 16:00:46] @ Quit: themurph: Quit: themurph | ||
| [2010/02/08 16:00:58] <elasticdog> jrojas: know of any reason why it would be failing without specifying those headers? | ||
| [2010/02/08 16:01:23] <jrojas> elasticdog: because apache doesnt know to pass the ssl info to puppet? | ||
| [2010/02/08 16:01:41] @ Quit: ckauhaus: Quit: Leaving. | ||
| [2010/02/08 16:01:55] <jrojas> the ssl verify happes @apache, but puppetmaster's rack app needs to know it as well. | ||
| [2010/02/08 16:01:57] @ Quit: tim|imac: Quit: tim|imac | ||
| [2010/02/08 16:02:17] <jrojas> well, thats my assumption at least | ||
| [2010/02/08 16:02:53] @ themurph joined channel #puppet | ||
| [2010/02/08 16:03:56] @ swygue joined channel #puppet | ||
| [2010/02/08 16:04:11] <elasticdog> jrojas: gotcha...perhaps that needs to be added into the provided template within the puppet gem | ||
| [2010/02/08 16:04:21] <jrojas> elasticdog: is it not? | ||
| [2010/02/08 16:04:50] <jrojas> ./ext/rack/files/apache2.conf | ||
| [2010/02/08 16:04:53] <jrojas> that has it | ||
| [2010/02/08 16:05:02] @ Quit: bug: Quit: bug | ||
| [2010/02/08 16:05:38] <jrojas> mine does at least.. | ||
| [2010/02/08 16:06:29] <elasticdog> jrojas: not there in the /puppet-0.25.4 gem sources, so it must have been fixed | ||
| [2010/02/08 16:08:02] <jrojas> odd maybe my config in my rpm was taked from the wiki or its the 24.x config.. | ||
| [2010/02/08 16:08:03] <jrojas> hmm | ||
| [2010/02/08 16:08:39] <jrojas> elasticdog: are you sure you didnt have an issue with the auth.conf ? | ||
| [2010/02/08 16:09:18] <elasticdog> jrojas: looks like it has been removed in the git repo (http://github.com/reductivelabs/puppet/blob/master/ext/rack/files/apache2.conf) | ||
| [2010/02/08 16:09:37] <jrojas> yeah, i guess i have an older config | ||
| [2010/02/08 16:09:38] @ Quit: heydrick1: Quit: Leaving. | ||
| [2010/02/08 16:09:40] <elasticdog> jrojas: I don't have an auth.conf, so maybe that's why I was having an issue | ||
| [2010/02/08 16:09:45] <jrojas> hmm | ||
| [2010/02/08 16:09:57] @ heydrick joined channel #puppet | ||
| [2010/02/08 16:10:03] @ bug joined channel #puppet | ||
| [2010/02/08 16:10:33] @ ambroff joined channel #puppet | ||
| [2010/02/08 16:11:20] <elasticdog> jrojas: there's not git history of its removal...guess I'll just leave it in for now, but wonder if it should be reported somewhere | ||
| [2010/02/08 16:11:45] <jrojas> maybe.. | ||
| [2010/02/08 16:11:56] <jrojas> i would try pinging the mailing list | ||
| [2010/02/08 16:12:04] @ OpenMedia joined channel #puppet | ||
| [2010/02/08 16:12:21] <jrojas> or find christian on irc and ask | ||
| [2010/02/08 16:12:47] <elasticdog> is there info on auth.conf somewhere...I noticed when running without passenger that puppetmasterd says there's a missing auth.conf, so it assumes some sane defaults | ||
| [2010/02/08 16:13:10] @ Quit: swygue: Remote host closed the connection | ||
| [2010/02/08 16:13:15] <elasticdog> perhaps when things go through passenger it doesn't degrade so gracefully when no auth.conf is present | ||
| [2010/02/08 16:13:18] <jrojas> hm not sure. there is really no good doc to say this is what you need | ||
| [2010/02/08 16:15:01] @ swygue joined channel #puppet | ||
| [2010/02/08 16:16:16] @ tim|imac joined channel #puppet | ||
| [2010/02/08 16:16:35] @ cegeddin joined channel #puppet | ||
| [2010/02/08 16:17:04] @ cegeddin left channel #puppet () | ||
| [2010/02/08 16:21:07] @ Quit: mrpots: Ping timeout: 248 seconds | ||
| [2010/02/08 16:23:40] @ Quit: tim|imac: Quit: tim|imac | ||
| [2010/02/08 16:24:11] @ Quit: mvn071: Quit: Leaving | ||
| [2010/02/08 16:26:48] @ ccaum left channel #puppet () | ||
| [2010/02/08 16:27:07] @ ccaum joined channel #puppet | ||
| [2010/02/08 16:27:39] <dan__t> I'm bored. | ||
| [2010/02/08 16:27:45] <jrojas> work. | ||
| [2010/02/08 16:28:29] @ Quit: giskard: Remote host closed the connection | ||
| [2010/02/08 16:33:09] <benkevan> jrojas: Heck yeah there is a need.. if anyone runs puppet on say.. SLES11.. they'd use zypper (although yum is available, zypper is the primary package manager) | ||
| [2010/02/08 16:34:09] <jrojas> benkevan: there is only the one ticket for anything zypper related, you may want to log a feature request.. | ||
| [2010/02/08 16:34:34] <benkevan> jrojas: Guess I should :D .. | ||
| [2010/02/08 16:35:22] <jrojas> benkevan: actually. | ||
| [2010/02/08 16:35:41] <benkevan> uh oh.. that sounds like a scary "actually" | ||
| [2010/02/08 16:35:42] <jrojas> are zypper repo configs like yum repo configs? | ||
| [2010/02/08 16:36:08] <jrojas> in /etc/zypp/repos.d ? | ||
| [2010/02/08 16:36:17] <benkevan> jrojas: there are def similarities.. | ||
| [2010/02/08 16:36:33] <benkevan> although i'm sure you can't use a yum provider for zypper, since the argument passing is different.. | ||
| [2010/02/08 16:36:39] <jrojas> hm | ||
| [2010/02/08 16:37:01] <jrojas> i was looking at the yumrepo type and realized its fairly similar | ||
| [2010/02/08 16:37:25] @ yarihm joined channel #puppet | ||
| [2010/02/08 16:38:14] <benkevan> yes.. they are fairly similar.. just slight differences.. but how zypper uses them is quite different | ||
| [2010/02/08 16:38:44] <jrojas> how is kinda moot, its just a matter of defining it so puppet can create teh repo :) | ||
| [2010/02/08 16:41:08] @ Tuf8 joined channel #puppet | ||
| [2010/02/08 16:42:17] <Tuf8> hey all.. is opensolaris 0906 supported as a distro type in cobbler? | ||
| [2010/02/08 16:42:32] <jrojas> Tuf8: did you ask in #cobbler ? :) | ||
| [2010/02/08 16:42:37] @ kolla joined channel #puppet | ||
| [2010/02/08 16:42:42] @ mikerowehl joined channel #puppet | ||
| [2010/02/08 16:42:47] <Tuf8> hah crap wrong channel ! | ||
| [2010/02/08 16:42:49] <Tuf8> sorry | ||
| [2010/02/08 16:43:13] <murkk> are there ubuntu/debian packages for 0.25.4 around? | ||
| [2010/02/08 16:43:57] @ Mick27 joined channel #puppet | ||
| [2010/02/08 16:47:12] <jrojas> i believe 0.25.4 made it into LTD | ||
| [2010/02/08 16:47:29] <jrojas> http://packages.debian.org/sid/puppet looks like its still stuck at 0.25.3 for debian | ||
| [2010/02/08 16:48:14] <benkevan> How would you compare something like slack to puppet? | ||
| [2010/02/08 16:48:21] <jrojas> slack? | ||
| [2010/02/08 16:48:41] @ doubleukay joined channel #puppet | ||
| [2010/02/08 16:48:48] <benkevan> http://www.sundell.net/~alan/projects/slack/ | ||
| [2010/02/08 16:48:58] <jrojas> http://code.google.com/p/slack/ | ||
| [2010/02/08 16:49:08] <jrojas> i think they do it for you | ||
| [2010/02/08 16:49:13] <jrojas> "Basically, it's a glorified wrapper around rsync. " | ||
| [2010/02/08 16:49:18] <jrojas> slack that is | ||
| [2010/02/08 16:49:43] <jrojas> hasnt been touched sinceapril 2009 | ||
| [2010/02/08 16:49:51] <jrojas> guess tehy are slacking? | ||
| [2010/02/08 16:49:55] <benkevan> agreed.. I have a home made wrapper process in house around rsync which seems to be quite a bit like slack.. still trying to grasp the gains over slack puppet has.. | ||
| [2010/02/08 16:50:20] <benkevan> (before I spend too much time doing something new.. you know?) | ||
| [2010/02/08 16:50:36] @ Quit: cliff-hm: Ping timeout: 245 seconds | ||
| [2010/02/08 16:50:41] <jb> if slack is a rsync wrapper, puppet is much more :) | ||
| [2010/02/08 16:50:53] <duritong> much much much much more | ||
| [2010/02/08 16:50:53] <benkevan> jb: Like? | ||
| [2010/02/08 16:51:06] <jb> uhm, rsync "manages" files | ||
| [2010/02/08 16:51:10] <duritong> installing packaging, managing users, triggering actions | ||
| [2010/02/08 16:51:10] @ dirtyalpaca joined channel #puppet | ||
| [2010/02/08 16:51:14] <duritong> and so on | ||
| [2010/02/08 16:51:16] <jb> puppet manages files, services, packages, etc, etc | ||
| [2010/02/08 16:51:23] @ Quit: doubleukay: Client Quit | ||
| [2010/02/08 16:51:31] <duritong> and the model behind the idea is basically different | ||
| [2010/02/08 16:51:44] <benkevan> wrapper script around rsync can do that.. but also have built in to check permissions of files.. changes to xxx configfiles etc etc.. | ||
| [2010/02/08 16:51:49] <jb> no it cant | ||
| [2010/02/08 16:52:11] <jb> well, I suppose it could very very sloppiliy | ||
| [2010/02/08 16:52:13] <jb> in some cases | ||
| [2010/02/08 16:52:20] <Volcane> user{"foo": ensure => present} | ||
| [2010/02/08 16:52:22] <Volcane> makes a user | ||
| [2010/02/08 16:52:23] <benkevan> jb: Mine does.. rsync is just the pull trigger.. it'll pull direction on what to do to the servers afterwards.. | ||
| [2010/02/08 16:52:26] <Volcane> on any of 10s of OS | ||
| [2010/02/08 16:52:30] <jb> exactly | ||
| [2010/02/08 16:52:33] <jb> any supported OS.. | ||
| [2010/02/08 16:52:37] <tessier> Ah....it is a good feeling when things actually work as you expect. | ||
| [2010/02/08 16:52:43] <Volcane> service{"foo": ensure => running} | ||
| [2010/02/08 16:52:45] <tessier> Makes me think I might be starting to understand this stuff. | ||
| [2010/02/08 16:52:46] <jb> your rsync cannot do that without a million hooks, and execs | ||
| [2010/02/08 16:53:06] <Volcane> file{"/etc/foo": content => template("some-dynamic-template.erb"), notify => Service["foo"]} | ||
| [2010/02/08 16:53:10] <Volcane> builds /etc/foo from a template | ||
| [2010/02/08 16:53:14] <jb> package "blah" { ensure => latest } | ||
| [2010/02/08 16:53:16] <Volcane> restarts the service should it ver change | ||
| [2010/02/08 16:53:28] <jb> ensures that the latest version of "blah" is installed | ||
| [2010/02/08 16:53:48] @ ckauhaus joined channel #puppet | ||
| [2010/02/08 16:53:56] <Volcane> so its a bit more advanced than your average rsync wrapper | ||
| [2010/02/08 16:54:00] <benkevan> my script will do that also.. does a curl of a version file of each package.. if update.. and it's said to upgrade then it upgrades.. | ||
| [2010/02/08 16:54:10] @ doubleukay joined channel #puppet | ||
| [2010/02/08 16:54:23] <benkevan> maybe I just need to really start using puppet to understand how much more it can actually do .. | ||
| [2010/02/08 16:54:30] <jb> maybe you should read the docs :) | ||
| [2010/02/08 16:54:34] <benkevan> just trying to justify the extra time I'll spend on the learning curve | ||
| [2010/02/08 16:54:35] <Volcane> benkevan: does your script have a book, training, commercial support and so forth for your next employee? | ||
| [2010/02/08 16:54:40] <jrojas> benkevan: package { "ruby": ensure => "1.9" } . . | ||
| [2010/02/08 16:54:47] <tessier> benkevan: The learning curve is indeed the isuse. | ||
| [2010/02/08 16:54:48] <tessier> issue | ||
| [2010/02/08 16:54:54] <benkevan> Volcane: Exactly.. that's why I've been working on going to a "standard" app :) .. | ||
| [2010/02/08 16:55:05] <tessier> benkevan: It's no worse than cfengine. But it sucks having to learn a second system. In the end I think it will be worth it though. | ||
| [2010/02/08 16:55:21] @ Quit: doubleukay: Client Quit | ||
| [2010/02/08 16:55:24] <Volcane> benkevan: also lots of consultants working in the space who you can just apy to come do it for you and tools like theforeman.org that integrates with it | ||
| [2010/02/08 16:55:25] <benkevan> tessier: Well.. right now I'm just on home made script.. with comments to hell | ||
| [2010/02/08 16:55:27] <jb> i really don't think the learning curve is that bad.. especially for basic things. | ||
| [2010/02/08 16:55:56] <benkevan> jb: ok.. well here I go.. (actually just installed the puppetmaster and going to set the puppetmastser as a puppet to test) | ||
| [2010/02/08 16:56:21] <jb> you will want to read up on module organization, environments, etc | ||
| [2010/02/08 16:56:27] <jb> if you want to be serious about it | ||
| [2010/02/08 16:56:58] <Volcane> benkevan: the learning curve is mostly related to the fact that its a whole different paradigm of programming - it is not functional programming, doesnt have loops etc, so u need to forget about shall scripts and how you used to do it and trying to recreate that, u need to start with an open mind | ||
| [2010/02/08 16:57:31] <benkevan> kinda dissapointed they don't have a zypper provider yet though :D .. .. is there also a way puppet can just trigger a local script? ie I can script up a zypper wrapper that puppet can call until puppet 25.5 is releaseed? | ||
| [2010/02/08 16:57:50] <Volcane> benkevan: you can exec stuff yes | ||
| [2010/02/08 16:58:17] <benkevan> Volcane: Agreed.. I am starting with an open mind.. I think even considering finding a more known central configuration tool as opposed to home made tools is a good step (i'm looking at slack / cfengine / puppet ) right now.. :) | ||
| [2010/02/08 16:58:19] @ Quit: swygue: Quit: Ex-Chat | ||
| [2010/02/08 16:59:19] @ doubleukay joined channel #puppet | ||
| [2010/02/08 16:59:37] <tessier> The documentation is great for the most part but I'm trying to figure out if I can use a case wrapped around a source => in a file type specification or if I have to define a $source outside of the file specification and I'm only seeing examples of case used to include classes like here: http://www.reductivelabs.com/trac/puppet/wiki/LanguageTutorial#case-statement | ||
| [2010/02/08 16:59:55] <tessier> benkevan: The only thing other than puppet that I might consider right now is Chef but puppet seems ahead in maturity and popularity. | ||
| [2010/02/08 17:00:15] <benkevan> tessier: Link? | ||
| [2010/02/08 17:00:43] <Volcane> tessier: what are you trying to do? did you look at selectors? | ||
| [2010/02/08 17:01:13] <tessier> benkevan: I think this is it here: http://wiki.opscode.com/display/chef/Home | ||
| [2010/02/08 17:01:29] <rmiller4pi8> tessier: yeah, selectors do what you are trying to do | ||
| [2010/02/08 17:01:36] <tessier> Volcane: I want to install a certain sudoers file on this machine, a different one on that machine, and a default everywhere else. | ||
| [2010/02/08 17:02:02] <jb> ah | ||
| [2010/02/08 17:02:03] <benkevan> http://bitfieldconsulting.com/puppet-vs-chef < Pretty much sums it up haha | ||
| [2010/02/08 17:02:03] <jb> extdata :) | ||
| [2010/02/08 17:02:05] @ Quit: ahasenack: Quit: Leaving | ||
| [2010/02/08 17:02:11] <tessier> ah, ok | ||
| [2010/02/08 17:02:13] <tessier> selectors it is. Thanks | ||
| [2010/02/08 17:02:16] <jb> well, thats how I would do it | ||
| [2010/02/08 17:02:22] <Volcane> tessier: source => ["puppet://.....", "puppet://....", ...] | ||
| [2010/02/08 17:02:27] <Volcane> tessier: the first one it finds it installs | ||
| [2010/02/08 17:02:42] <tessier> Volcane: Why wouldn't it always find the first one? | ||
| [2010/02/08 17:02:58] <Volcane> tessier: "puppet:///path/to/file.${fqdn}" | ||
| [2010/02/08 17:03:02] <benkevan> damn .. puppets site is taking up nearly 1/2 my tabs in my browser haha.. any good puppet books you guys would recommend? | ||
| [2010/02/08 17:03:04] <tessier> Volcane: Ah. | ||
| [2010/02/08 17:03:14] @ Quit: doubleukay: Quit: ZNC - http://znc.sourceforge.net | ||
| [2010/02/08 17:03:25] <Volcane> tessier: so you have allow file overrides for host, domain, role, client, etc whatever u want | ||
| [2010/02/08 17:03:31] <tessier> Volcane: And is the last one on the list a default? | ||
| [2010/02/08 17:03:45] <Volcane> tessier: yes | ||
| [2010/02/08 17:03:59] <tessier> benkevan: there is only one good puppet book. The one I have open in front of me right now...Pulling Strings With Puppet | ||
| [2010/02/08 17:04:04] <jb> benkevan: http://www.amazon.com/Pulling-Strings-Puppet-Configuration-Management/dp/1590599780 | ||
| [2010/02/08 17:04:08] <tessier> benkevan: By our very own jturnbull | ||
| [2010/02/08 17:04:14] <jb> although, I think the second edition is due soon | ||
| [2010/02/08 17:04:26] <tessier> Second edition? Sweet. I have noticed a few errors in this one. | ||
| [2010/02/08 17:04:30] <tessier> I'll have to pick it up. | ||
| [2010/02/08 17:05:00] <jb> excellent start though | ||
| [2010/02/08 17:05:04] <jb> and a very easy read | ||
| [2010/02/08 17:05:11] <tessier> I don't know about easy read... | ||
| [2010/02/08 17:05:16] <jb> well, I thought it was :) | ||
| [2010/02/08 17:05:22] <tessier> I have had to read it a couple of times and I still don't grok everything. Just a lot of information. | ||
| [2010/02/08 17:05:39] <tessier> It also bugs me how he shows how to do things this way but then says this other way is actually better etc. | ||
| [2010/02/08 17:06:22] <tessier> At one point he does import "groups/*" but never mentions that we should have created a groups directory in the puppet root to put the virt_groups.pp file we just created into etc. | ||
| [2010/02/08 17:06:42] <tessier> Hopefully he expanded things a bit and was able to talk them into providing an index at the end of the book. | ||
| [2010/02/08 17:06:42] <jb> yeah, my manifest organization is completely different | ||
| [2010/02/08 17:06:49] <jb> i use modules for everythingg | ||
| [2010/02/08 17:06:50] <benkevan> also how detremental are upgrades of puppet usually? | ||
| [2010/02/08 17:06:54] <tessier> I use modules for almost everything. | ||
| [2010/02/08 17:06:57] <benkevan> like upgrading from nagios 2 to nagios 3? | ||
| [2010/02/08 17:07:14] <jb> benkevan: i started at 0.25.1, and have moved to .4 without any issues | ||
| [2010/02/08 17:07:23] <jb> rolled out RPMS, and nothing else really.. | ||
| [2010/02/08 17:07:43] <benkevan> jb: using yum? I'm wondering how the prodivder move from yum to zypper would be once implemented.. | ||
| [2010/02/08 17:07:54] <jb> yes, I use yum | ||
| [2010/02/08 17:08:13] <jb> and either my custom deployment methods to push out to all machines/groups of machines | ||
| [2010/02/08 17:08:31] <jb> or just use puppet (package "puppet"{ ensure => "0.25.4" }) | ||
| [2010/02/08 17:08:44] @ doubleukay joined channel #puppet | ||
| [2010/02/08 17:11:09] <kami`> jrojas: thank you for the inline_template hint. It did the trick. | ||
| [2010/02/08 17:11:13] <kami`> Good night | ||
| [2010/02/08 17:11:17] @ kami` left channel #puppet ("ERC Version 5.3 (IRC client for Emacs)") | ||
| [2010/02/08 17:11:37] @ Djelibeybi joined channel #puppet | ||
| [2010/02/08 17:11:52] @ jfreeman joined channel #puppet | ||
| [2010/02/08 17:13:14] @ Quit: bug: Quit: bug | ||
| [2010/02/08 17:15:07] @ Quit: brothers: Quit: brothers | ||
| [2010/02/08 17:15:12] @ brothers joined channel #puppet | ||
| [2010/02/08 17:15:16] @ Quit: tyll: Quit: leaving | ||
| [2010/02/08 17:21:07] <benkevan> is there any special dns naming that is "required" for the puppetmaster? | ||
| [2010/02/08 17:21:20] <benkevan> or just defined in /etc/sysconfig/puppet | ||
| [2010/02/08 17:21:50] <benkevan> Reason I ask is because I just started puppetmaster with basic classes and get: Starting puppetmaster: dnsdomainname: Unknown host | ||
| [2010/02/08 17:22:00] <benkevan> (note that's the puppet master starting) .. | ||
| [2010/02/08 17:24:51] @ Quit: ckauhaus: Quit: Leaving. | ||
| [2010/02/08 17:28:11] <eric0> benkevan: its preferable to have a name 'puppet' pointed at your puppetmaster , because then you don't have to specify --server on the clients. | ||
| [2010/02/08 17:28:18] @ giskard joined channel #puppet | ||
| [2010/02/08 17:29:17] @ Quit: themurph: Quit: themurph | ||
| [2010/02/08 17:29:18] <Djelibeybi> benkevan: check that your DNS and /etc/hosts is properly configured. Do you get a sensible output when you run dnsdomainname manually? | ||
| [2010/02/08 17:29:55] <Djelibeybi> Also, while you can set the puppet FQDN in /etc/sysconfig/puppet, it is only used by the init.d script. Rather, if you don't want to use the puppet alias, set the FQDN in /etc/puppet/puppet.conf so that it is used by both the init.d script and manual puppetd runs | ||
| [2010/02/08 17:30:49] <benkevan> Yeah.. I just switched boxes.. and forgot I didn't have dns.. setup yet.. haha (this is just a test puppet for a poc).. so I just added to hosts.. sorry .. figured it out shortly after I posted that dnsdomainname came back neg | ||
| [2010/02/08 17:31:50] @ unxfrek joined channel #puppet | ||
| [2010/02/08 17:32:17] <benkevan> so this may sound like a stupid question.. when does the puppet master decide when to control the puppets? ie check the configs etc (you have to remember I come from a rsync pull process) .. :) | ||
| [2010/02/08 17:32:28] @ doubleukay left channel #puppet () | ||
| [2010/02/08 17:33:03] @ doubleukay joined channel #puppet | ||
| [2010/02/08 17:33:04] <eric0> clients wake up based on their schedule (default 30 min) and check in | ||
| [2010/02/08 17:33:46] <eric0> thats if you have puppetd running as a persistent daemon, though seems like lots of people run 'puppetd --onetime' out of cron instead | ||
| [2010/02/08 17:33:58] <Djelibeybi> benkevan: clients decide, not the puppetmaster. They poll every 30 minutes by default | ||
| [2010/02/08 17:35:11] <benkevan> eric0: Djelibeybi: Thanks.. yeah I'm thinking of doing in cron instead.. (for controlled change control instances maybe).. where is the polling defined? just in /etc/sysconfig/puppet? (sorry for the basic questions.. i know I should just RTFM, but I just setup so getting my toes wet) | ||
| [2010/02/08 17:35:31] <Djelibeybi> benkevan: do it in /etc/puppet/puppet.conf | ||
| [2010/02/08 17:35:40] <Djelibeybi> benkevan: or don't, if you're going to use cron | ||
| [2010/02/08 17:36:00] <Djelibeybi> benkevan: or, if you want to be able to manually trigger, look into puppetrun instead (where the puppet client listens for run requests from the master) | ||
| [2010/02/08 17:36:33] <doubleukay> eric0: what would be the dis/advantages of running puppetd as a corn? | ||
| [2010/02/08 17:36:38] <benkevan> wish the default puppet.conf had more options in there (just commented out) so you knoew where they would go by default :) | ||
| [2010/02/08 17:36:50] <dirtyalpaca> hey guys... i was wondering if anyone knows what this might mean... http://pastie.org/815393 | ||
| [2010/02/08 17:36:54] <benkevan> doubleukay: I think control would be the biggest one.. | ||
| [2010/02/08 17:37:02] @ Quit: paxos: Quit: paxos | ||
| [2010/02/08 17:37:17] <nevyn> doubleukay: I'm not sure about the computational power of corn. | ||
| [2010/02/08 17:38:06] <Djelibeybi> benkevan: no, more like memory management. The ruby that ships with EL4/EL5 leaks like a rusty sieve. | ||
| [2010/02/08 17:38:34] <Djelibeybi> And, it means puppetd is just sitting running all the time. | ||
| [2010/02/08 17:38:47] <nevyn> Djelibeybi: rusty sieves leak less than non-rusty ones | ||
| [2010/02/08 17:38:56] <doubleukay> true that | ||
| [2010/02/08 17:39:00] <nevyn> as the rust closes up the sieve holes... | ||
| [2010/02/08 17:39:11] <Djelibeybi> nevyn: now is not the time for logic. | ||
| [2010/02/08 17:39:26] <Djelibeybi> nevyn: but, the rust could also eat into the sieve and make larger holes. | ||
| [2010/02/08 17:39:29] @ Quit: rgsteele: Remote host closed the connection | ||
| [2010/02/08 17:39:33] <eric0> if you want to trigger execution via puppetrun it obviously needs to run as persistent daemon | ||
| [2010/02/08 17:40:22] <nevyn> Djelibeybi: well at the point it's no longer recognisable as a sieve I'd contend (but I am being silly mostly) | ||
| [2010/02/08 17:40:34] @ Quit: allsystemsarego: Quit: Leaving | ||
| [2010/02/08 17:40:38] <Djelibeybi> nevyn: it's a Worlde Gonne Madde! | ||
| [2010/02/08 17:41:20] @ bobbyz joined channel #puppet | ||
| [2010/02/08 17:43:13] @ thomas54321 joined channel #puppet | ||
| [2010/02/08 17:43:19] <benkevan> yay.. did my first test with sudoers .. haha | ||
| [2010/02/08 17:43:37] <thomas54321> hi - i'm trying to distribute custom facts using the wiki documentation (http://reductivelabs.com/trac/puppet/wiki/PluginsInModules) - could anyone help? | ||
| [2010/02/08 17:44:18] <thomas54321> as far as I understand, inside the module, you create a lib/facter/fact.rb file with the fact | ||
| [2010/02/08 17:44:43] <thomas54321> and add pluginsync and factpath to puppet.conf | ||
| [2010/02/08 17:44:46] <thomas54321> am I missing something ? | ||
| [2010/02/08 17:47:23] @ Quit: unxfrek: Remote host closed the connection | ||
| [2010/02/08 17:47:33] @ plathrop joined channel #puppet | ||
| [2010/02/08 17:48:07] <shenson> thomas54321, running what version? | ||
| [2010/02/08 17:48:14] <thomas54321> latest | ||
| [2010/02/08 17:48:33] <shenson> I've found that I need to run puppet on the puppet master before the facts and plugins are seen | ||
| [2010/02/08 17:49:00] <thomas54321> okay, so i'm not missing anything ? | ||
| [2010/02/08 17:49:07] <shenson> not sure if that is an artifact of the way I have puppet setup, but it works for me | ||
| [2010/02/08 17:49:33] <shenson> in my setup puppet on the puppet master syncs the facts into the place that the puppetmasterd process looks for them and then it syncs them out to clients | ||
| [2010/02/08 17:49:35] <thomas54321> how can i test that the fact is working properly? should i show up on the clients ? | ||
| [2010/02/08 17:50:14] <shenson> set FACTERLIB to where they sync down on the client (/var/lib/puppet/{plugins,lib,facts}) and then run facter | ||
| [2010/02/08 17:50:17] <thomas54321> the facts show up, i mean | ||
| [2010/02/08 17:50:51] <thomas54321> ah, ok | ||
| [2010/02/08 17:50:53] <thomas54321> thanks! | ||
| [2010/02/08 17:50:56] <shenson> np | ||
| [2010/02/08 17:50:59] @ Quit: thomas54321: Quit: thomas54321 | ||
| [2010/02/08 17:52:07] @ Quit: plathrop: Client Quit | ||
| [2010/02/08 17:52:22] @ plathrop joined channel #puppet | ||
| [2010/02/08 17:53:05] @ Quit: verwilst: Quit: Ex-Chat | ||
| [2010/02/08 17:53:32] @ akhet joined channel #puppet | ||
| [2010/02/08 17:55:43] @ bug joined channel #puppet | ||
| [2010/02/08 18:00:28] @ ccaum left channel #puppet () | ||
| [2010/02/08 18:04:45] @ Quit: pheezy: Remote host closed the connection | ||
| [2010/02/08 18:05:49] @ Quit: malraid: Ping timeout: 264 seconds | ||
| [2010/02/08 18:07:26] @ doubleukay left channel #puppet () | ||
| [2010/02/08 18:08:12] @ fdz joined channel #puppet | ||
| [2010/02/08 18:08:22] @ doubleukay joined channel #puppet | ||
| [2010/02/08 18:13:07] @ Quit: fcoury: Quit: <original quit message here> | ||
| [2010/02/08 18:14:30] @ Quit: fdz: Quit: Leaving. | ||
| [2010/02/08 18:15:07] @ doubleukay left channel #puppet ("Leaving") | ||
| [2010/02/08 18:16:55] @ pheezy joined channel #puppet | ||
| [2010/02/08 18:17:42] @ doubleukay joined channel #puppet | ||
| [2010/02/08 18:18:18] @ themurph joined channel #puppet | ||
| [2010/02/08 18:19:21] @ svend joined channel #puppet | ||
| [2010/02/08 18:20:40] <svend> I'm trying to get puppetrun to work. I have the puppetd listening and namespaceauth.conf set up. I don't see anything happening on the client when I run puppetrun. | ||
| [2010/02/08 18:21:16] <svend> I can connect to 8139 on the client from the server using openssl s_client. | ||
| [2010/02/08 18:21:55] @ blentz joined channel #puppet | ||
| [2010/02/08 18:22:59] @ blentz is now known as wakko666 | ||
| [2010/02/08 18:30:06] <Volcane> do you see tcp packets going from server to server? | ||
| [2010/02/08 18:30:41] @ Quit: vzctl_: Read error: Operation timed out | ||
| [2010/02/08 18:33:40] @ Quit: tonyskapunk: Remote host closed the connection | ||
| [2010/02/08 18:34:03] @ tonyskapunk joined channel #puppet | ||
| [2010/02/08 18:34:44] <svend> No, there are no packets on the client | ||
| [2010/02/08 18:34:50] @ Quit: tonyskapunk: Remote host closed the connection | ||
| [2010/02/08 18:35:13] @ tonyskapunk joined channel #puppet | ||
| [2010/02/08 18:35:31] <Volcane> what version? | ||
| [2010/02/08 18:37:37] <Volcane> there used to be a lame thing where it would ping the server first and if that failed it wont try to connect | ||
| [2010/02/08 18:39:39] @ Quit: bug: Quit: bug | ||
| [2010/02/08 18:39:41] @ Quit: jcape: Ping timeout: 265 seconds | ||
| [2010/02/08 18:41:10] @ Quit: autowidget: Quit: autowidget | ||
| [2010/02/08 18:41:39] <nevyn> cause nobody blocks ICMP echo internally | ||
| [2010/02/08 18:41:58] <nevyn> grumble grumble stupid networks people grumble | ||
| [2010/02/08 18:43:06] @ \ask joined channel #puppet | ||
| [2010/02/08 18:43:54] @ Quit: Ramonster: Quit: leaving | ||
| [2010/02/08 18:43:54] * Boney waves to nevyn | ||
| [2010/02/08 18:45:24] <svend> Volcane: puppet 0.24.5 | ||
| [2010/02/08 18:45:35] <Volcane> do you allow icmp? | ||
| [2010/02/08 18:45:44] <nevyn> Boney: hi | ||
| [2010/02/08 18:45:48] <Volcane> tcpdump on the box you run puppetrun on and see if anything goes to the target | ||
| [2010/02/08 18:45:48] @ vzctl_ joined channel #puppet | ||
| [2010/02/08 18:46:17] <z00dax> humm we were talking about icmp stuff a few days back | ||
| [2010/02/08 18:46:40] @ Quit: themurph: Quit: themurph | ||
| [2010/02/08 18:47:15] <tessier> Do most people also run puppetd from a cron to ensure that puppetd gets restarted in case it somehow dies but cron survives? | ||
| [2010/02/08 18:47:53] <nevyn> tessier: I have a cron resource which restarts puppet every day | ||
| [2010/02/08 18:48:06] * Volcane doesnt run the daemon | ||
| [2010/02/08 18:48:11] <Volcane> too many mem leaks and shit | ||
| [2010/02/08 18:48:22] <tessier> Yikes. | ||
| [2010/02/08 18:48:43] <z00dax> I've got puppet running from cron everywhere | ||
| [2010/02/08 18:49:29] <dan__t> yep. | ||
| [2010/02/08 18:49:46] <dan__t> I wish our VPN wasn't a piece of shit. | ||
| [2010/02/08 18:49:47] <tessier> z00dax: If your cron dies then what? | ||
| [2010/02/08 18:49:54] <tessier> dan__t: openvpn ftw | ||
| [2010/02/08 18:50:18] <dan__t> i have like $50k of vpn terminators out there | ||
| [2010/02/08 18:50:20] <dan__t> they all suck. | ||
| [2010/02/08 18:50:37] <plathrop> If I want to use a generate script to generate a list of resource names, how do I get puppet to see the output of generate as a comma separated list instead of one giant string? | ||
| [2010/02/08 18:50:44] <dan__t> I take that back. Its the vpnc networkmanager plugin that sucks. | ||
| [2010/02/08 18:50:49] <plathrop> ugh. That's a crappy question... let me pastie | ||
| [2010/02/08 18:51:10] <danbeck> issue tracker failing? http://projects.reductivelabs.com/ | ||
| [2010/02/08 18:51:14] <danbeck> I get a 404 | ||
| [2010/02/08 18:51:54] @ Quit: rmiller4pi8: Ping timeout: 272 seconds | ||
| [2010/02/08 18:52:27] <z00dax> tessier: which is one of the reasons why nagios monitors last-puppet-run-age, and warns if > 6 hrs | ||
| [2010/02/08 18:52:45] <svend> I got it to work by modifying the puppetrun script. The hosts array gets truncated after Puppet.parse_config | ||
| [2010/02/08 18:52:47] <dan__t> i have a similar check with zabbix | ||
| [2010/02/08 18:52:50] <plathrop> http://pastie.org/815495 is the basic idea. | ||
| [2010/02/08 18:52:55] <dan__t> because nagios is the bastard child of everything unholy. | ||
| [2010/02/08 18:52:59] <z00dax> on most machines, cron tends to be involved with slightly more important things as well - like production things, which in turn are also monitored using nagios | ||
| [2010/02/08 18:53:04] <danbeck> yeah, http://projects.reductivelabs.com/ is down... | ||
| [2010/02/08 18:53:09] <plathrop> I'm doing this, and puppet is seeing $generated_names as one string instead of an array. | ||
| [2010/02/08 18:54:06] <plathrop> Oh, duh. | ||
| [2010/02/08 18:54:42] <plathrop> nevermind, figured it out | ||
| [2010/02/08 18:54:45] <tessier> plathrop: I love how explaining the problem to someone else often reveals the answer. :) | ||
| [2010/02/08 18:55:17] <tessier> Fully half of the problems I typically explain on IRC end up resulting in an "oh!" right after I finish explaining. | ||
| [2010/02/08 18:55:48] <tessier> But I always thank the denizens for lisetning. :) | ||
| [2010/02/08 18:55:52] <tessier> listening | ||
| [2010/02/08 18:55:59] <plathrop> Yes, thanks for listening | ||
| [2010/02/08 18:56:11] @ Quit: danbeck: Quit: danbeck | ||
| [2010/02/08 18:57:02] <zirpu> what was the answer? :-) | ||
| [2010/02/08 18:57:25] <plathrop> the "split" function. | ||
| [2010/02/08 19:01:18] @ Quit: dirtyalpaca: Quit: Leaving. | ||
| [2010/02/08 19:02:39] <tessier> Hmm...I want to ensure crond is running and the appropriate package is installed. So I have a service { crond: and a require => Package["vixie-cron"] which appears to be the name of the crond package in CentOS. Yet puppetd says Configuration could not be instantiated: Could not find dependency Package[vixie-cron] for Service[crond] | ||
| [2010/02/08 19:03:01] <tessier> yum search crond shows vixie-cron is available... | ||
| [2010/02/08 19:03:06] <tessier> er..yum search cron | ||
| [2010/02/08 19:03:24] <dan__t> Its vixie-cron | ||
| [2010/02/08 19:07:59] <z00dax> tessier: did you define Package vixie-cron somewhere first ? | ||
| [2010/02/08 19:08:08] <tessier> dan__t: Right...and that is what I specified for the Package: require => Package["vixie-cron"] | ||
| [2010/02/08 19:08:17] <tessier> z00dax: No...I thought it would automatically pull it in with yum. | ||
| [2010/02/08 19:08:35] <z00dax> tessier: puppet isnt yum | ||
| [2010/02/08 19:11:31] <tessier> I got it. Thanks. | ||
| [2010/02/08 19:12:56] @ Quit: ezmobius: Remote host closed the connection | ||
| [2010/02/08 19:14:58] @ Quit: jab_doa: Quit: Verlassend | ||
| [2010/02/08 19:18:53] @ Quit: giskard: Remote host closed the connection | ||
| [2010/02/08 19:20:04] @ Quit: shenson: Quit: /me taps out | ||
| [2010/02/08 19:22:06] @ Quit: jes5: Ping timeout: 276 seconds | ||
| [2010/02/08 19:24:43] @ thomas54321 joined channel #puppet | ||
| [2010/02/08 19:28:15] @ pinoyskull joined channel #puppet | ||
| [2010/02/08 19:29:24] @ Quit: benkevan: Remote host closed the connection | ||
| [2010/02/08 19:31:33] @ Quit: pheezy: Remote host closed the connection | ||
| [2010/02/08 19:34:49] <dan__t> Good. So what's the problem? | ||
| [2010/02/08 19:35:53] @ malraid joined channel #puppet | ||
| [2010/02/08 19:36:02] @ Quit: malraid: Client Quit | ||
| [2010/02/08 19:39:08] @ Quit: Risar: Quit: Coyote finally caught me | ||
| [2010/02/08 19:41:06] @ Quit: lak: Quit: lak | ||
| [2010/02/08 19:45:36] @ Quit: brothers: Ping timeout: 245 seconds | ||
| [2010/02/08 19:45:55] @ Quit: kaptk2: Quit: Leaving. | ||
| [2010/02/08 19:47:44] @ rmiller4pi8 joined channel #puppet | ||
| [2010/02/08 19:48:56] @ bodepd joined channel #puppet | ||
| [2010/02/08 19:49:01] @ Quit: lutter: Ping timeout: 264 seconds | ||
| [2010/02/08 19:53:55] @ Quit: bodepd: Quit: bodepd | ||
| [2010/02/08 19:54:21] @ Quit: pinoyskull: Ping timeout: 245 seconds | ||
| [2010/02/08 19:54:53] @ Quit: Mick27: Quit: This computer has gone to sleep | ||
| [2010/02/08 19:57:55] @ Quit: rmiller4pi8: Ping timeout: 240 seconds | ||
| [2010/02/08 19:59:26] @ Quit: warreng: Quit: warreng | ||
| [2010/02/08 20:00:15] @ ghg joined channel #puppet | ||
| [2010/02/08 20:00:42] @ Quit: qwebirc99364: Disconnected by services | ||
| [2010/02/08 20:00:48] @ Quit: ghg: Client Quit | ||
| [2010/02/08 20:06:55] @ Quit: akhet: Quit: l8z | ||
| [2010/02/08 20:07:33] @ Risar joined channel #puppet | ||
| [2010/02/08 20:13:37] @ Quit: jaredrhine: Ping timeout: 272 seconds | ||
| [2010/02/08 20:16:43] @ quit (okay bye) | ||
| [2010/02/08 20:18:06] @ Joined channel #puppet | ||
| [2010/02/08 20:18:06] @ Topic is "Puppet 0.25.4 released - http://bit.ly/9c5HMv | Log bugs and feature requests at http://tinyurl.com/46ryk7 | see also #puppet-dev" | ||
| [2010/02/08 20:18:06] @ Topic set by mpdehaan!~mpdehaan@cpe-069-134-171-066.nc.res.rr.com on Tue Feb 02 17:04:30 -0500 2010 | ||
| [2010/02/08 20:18:19] @ Mode +cnt by anthony.freenode.net | ||
| [2010/02/08 20:20:08] @ pheezy joined channel #puppet | ||
| [2010/02/08 20:24:34] @ Quit: himanshu: Remote host closed the connection | ||
| [2010/02/08 20:25:08] @ warreng joined channel #puppet | ||
| [2010/02/08 20:26:49] @ Quit: eshamow: Quit: eshamow | ||
| [2010/02/08 20:27:51] <thomas54321> does anyone have experience adding a custom fact via a module? i still can't get it to work | ||
| [2010/02/08 20:28:55] <ohadlevy> thomas54321: does it get synced? | ||
| [2010/02/08 20:29:03] <thomas54321> no.. | ||
| [2010/02/08 20:29:13] <thomas54321> facter with FACTERLIB works | ||
| [2010/02/08 20:29:30] <thomas54321> puppet.conf has this under [main] | ||
| [2010/02/08 20:29:30] <thomas54321> pluginsync = true | ||
| [2010/02/08 20:29:31] <thomas54321> factpath = $vardir/lib/facter | ||
| [2010/02/08 20:29:35] <thomas54321> pluginsync = true | ||
| [2010/02/08 20:29:35] <thomas54321> factpath = $vardir/lib/facter | ||
| [2010/02/08 20:29:36] <ohadlevy> thomas54321: 1. do you have pluginsync enabled? 2. are you putting the file in the right place? | ||
| [2010/02/08 20:29:42] <thomas54321> pluginsync = true | ||
| [2010/02/08 20:29:46] <thomas54321> factpath = $vardir/lib/facter | ||
| [2010/02/08 20:30:10] <thomas54321> the fact is in /etc/puppet/modules/custom/lib/facter/fact.rb | ||
| [2010/02/08 20:30:59] <ohadlevy> permissions? | ||
| [2010/02/08 20:31:25] <thomas54321> world readable - do i need to add any flags to puppetd on the client to get the facts ? | ||
| [2010/02/08 20:32:11] <ohadlevy> the client need to have pluginsync=true | ||
| [2010/02/08 20:32:14] @ bodepd joined channel #puppet | ||
| [2010/02/08 20:33:23] <thomas54321> oh, i thought that was server side config | ||
| [2010/02/08 20:34:16] <thomas54321> well, that was definitely the problem | ||
| [2010/02/08 20:34:28] <thomas54321> thanks much... | ||
| [2010/02/08 20:35:10] @ jcape joined channel #puppet | ||
| [2010/02/08 20:35:36] @ Quit: tonyskapunk: Ping timeout: 240 seconds | ||
| [2010/02/08 20:36:56] @ Quit: bobbyz: Read error: Operation timed out | ||
| [2010/02/08 20:37:26] @ cliff-hm joined channel #puppet | ||
| [2010/02/08 20:39:18] @ bobbyz joined channel #puppet | ||
| [2010/02/08 20:40:25] @ Quit: murkk: Quit: murkk | ||
| [2010/02/08 20:40:47] <thomas54321> is there a way to enable pluginsync and factpath from the command line on the client ? | ||
| [2010/02/08 20:40:59] <ohadlevy> --pluginsync | ||
| [2010/02/08 20:41:57] <thomas54321> makes sense | ||
| [2010/02/08 20:44:10] @ Quit: yarihm: Quit: This computer has gone to sleep | ||
| [2010/02/08 20:45:41] @ Quit: Tuf8: Read error: Connection reset by peer | ||
| [2010/02/08 20:58:24] <slash^> do you guys find that you end up with alot of puppetmasterd procs that just seem to linger?? | ||
| [2010/02/08 20:58:40] <slash^> like a child is spawned but never gets cleaned up... | ||
| [2010/02/08 20:59:33] <slash^> just did a ps -auxww on my system and i have alot of them just sitting there as an idle interupted thread... | ||
| [2010/02/08 21:02:44] <slash^> as a temp fix ive moved to using passenger with apache | ||
| [2010/02/08 21:02:50] @ brothers joined channel #puppet | ||
| [2010/02/08 21:05:32] @ Quit: \ask: Ping timeout: 272 seconds | ||
| [2010/02/08 21:07:00] <ericlee> hi, how do I let puppet force to read the manifests? | ||
| [2010/02/08 21:07:17] <ericlee> I don't wanna restart the service every time. | ||
| [2010/02/08 21:10:20] @ Quit: svend: Ping timeout: 246 seconds | ||
| [2010/02/08 21:11:09] @ Quit: jfreeman: Remote host closed the connection | ||
| [2010/02/08 21:12:18] @ brhelwig joined channel #puppet | ||
| [2010/02/08 21:12:53] @ Quit: Risar: Remote host closed the connection | ||
| [2010/02/08 21:12:56] @ Risar joined channel #puppet | ||
| [2010/02/08 21:13:26] @ jfreeman joined channel #puppet | ||
| [2010/02/08 21:14:34] @ brhelwig left channel #puppet () | ||
| [2010/02/08 21:16:52] @ Quit: gebi: Read error: Operation timed out | ||
| [2010/02/08 21:36:19] @ bug joined channel #puppet | ||
| [2010/02/08 21:36:27] @ Djelibeybi joined channel #puppet | ||
| [2010/02/08 21:40:53] @ themurph joined channel #puppet | ||
| [2010/02/08 21:40:57] @ noyb joined channel #puppet | ||
| [2010/02/08 21:42:49] @ Quit: themurph: Client Quit | ||
| [2010/02/08 21:51:01] @ Quit: Welsh_Dwarf: Ping timeout: 245 seconds | ||
| [2010/02/08 21:52:58] @ Quit: pheezy: Remote host closed the connection | ||
| [2010/02/08 21:57:25] @ pinoyskull joined channel #puppet | ||
| [2010/02/08 22:02:26] <ericlee> I wrote a service class but when the client starts the service is not running. | ||
| [2010/02/08 22:02:43] <ericlee> I will paste my class. | ||
| [2010/02/08 22:03:12] <ericlee> http://pastebin.com/me1479db | ||
| [2010/02/08 22:03:14] <ericlee> Thanks. | ||
| [2010/02/08 22:04:41] @ Quit: warreng: Quit: warreng | ||
| [2010/02/08 22:10:05] <ericlee> Anyone could help me here? | ||
| [2010/02/08 22:20:23] @ pheezy joined channel #puppet | ||
| [2010/02/08 22:21:21] <bodepd> is it possible that you are not including the class? | ||
| [2010/02/08 22:26:12] @ murkk joined channel #puppet | ||
| [2010/02/08 22:28:37] <murkk> I get Duplicate parameter 'require' for on Exec | ||
| [2010/02/08 22:28:57] <murkk> I have two requires in the exec blog, can you not do that? | ||
| [2010/02/08 22:29:01] @ jaredrhine joined channel #puppet | ||
| [2010/02/08 22:29:35] <bodepd> use an array: require => [Service[1], Service[2]] | ||
| [2010/02/08 22:35:52] @ Quit: TheAvatar: Read error: Operation timed out | ||
| [2010/02/08 22:36:07] @ PaulWay[w] joined channel #puppet | ||
| [2010/02/08 22:37:21] <murkk> bodepd: thanks, in the past I have had two requires, is that something new? | ||
| [2010/02/08 22:37:27] @ TheAvatar joined channel #puppet | ||
| [2010/02/08 22:39:31] <bodepd> I know its always been true for 0.25.x | ||
| [2010/02/08 22:39:45] <tmz> As well for 0.24 AFAIK | ||
| [2010/02/08 22:40:01] <ohadlevy> murkk: 0.24.x would allow the syntax, but would ignore one of the required statements, 0.25.x report it as an error | ||
| [2010/02/08 22:40:19] <murkk> thanks guys | ||
| [2010/02/08 22:40:40] <bodepd> ohadlevy: your good! I was about to switch to the old version and test :) | ||
| [2010/02/08 22:41:24] <ohadlevy> bodepd: it was once of the first thing you notice when you try to run your 0.24x manifest on a 0.25 master :) | ||
| [2010/02/08 22:43:04] @ \ask joined channel #puppet | ||
| [2010/02/08 22:50:11] @ RageLink joined channel #puppet | ||
| [2010/02/08 23:00:21] @ Quit: murkk: Quit: murkk | ||
| [2010/02/08 23:10:37] @ plathrop is now known as plathrop-away | ||
| [2010/02/08 23:12:31] @ qwebirc77056 joined channel #puppet | ||
| [2010/02/08 23:14:06] @ Quit: cliff-hm: Ping timeout: 272 seconds | ||
| [2010/02/08 23:19:32] @ Quit: LapTop006: Ping timeout: 265 seconds | ||
| [2010/02/08 23:20:47] @ Quit: qwebirc77056: Quit: Page closed | ||
| [2010/02/08 23:21:39] @ Quit: pheezy: Remote host closed the connection | ||
| [2010/02/08 23:22:13] @ pheezy joined channel #puppet | ||
| [2010/02/08 23:29:48] <thomas54321> in what directory in my puppet module do i need to put custom types for puppet 0.25 ? | ||
| [2010/02/08 23:33:02] @ Quit: thomas54321: Quit: thomas54321 | ||
| [2010/02/08 23:40:47] @ ezmobius joined channel #puppet | ||
| [2010/02/08 23:47:18] @ Quit: justdave: Remote host closed the connection | ||
| [2010/02/08 23:47:38] @ Quit: noyb: Quit: noyb | ||
| [2010/02/08 23:54:23] @ rmiller4pi8 joined channel #puppet | ||
| [2010/02/08 23:55:29] @ Quit: ezmobius: Remote host closed the connection |
Generated by irclog2html.py 2.8 by Marius Gedminas - find it at mg.pov.lt!